2014 Doctor of Philosophy Thesis in Mechanical Engineering, Production Technology branch, supervised by Professor Altino de Jesus Roque Loureiro and Professor Albano Augusto Cavaleiro Rodrigues de Carvalho and submitted to the Mechanical Engineering Department of the Faculty of Sciences and Technology of the University of Coimbra. SURFACE MODIFICATION OF MOLDS AND ACESSORIES FOR THE GLASS INDUSTRY Filipe Daniel Fernandes IMAGEM
227
Embed
surface modification of molds and acessories for the glass industry
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
2014
Doctor of Philosophy Thesis in Mechanical Engineering, Production Technology branch, supervised by Professor Altino de Jesus
Roque Loureiro and Professor Albano Augusto Cavaleiro Rodrigues de Carvalho and submitted to the Mechanical Engineering
Department of the Faculty of Sciences and Technology of the University of Coimbra.
SURFACE MODIFICATION OF MOLDS AND ACESSORIES FOR THE GLASS INDUSTRY
Filipe Daniel Fernandes
IMAGEM
2014
Doctor of Philosophy Thesis in Mechanical Engineering, Production Technology branch, supervised by Professor Altino de Jesus
Roque Loureiro and Professor Albano Augusto Cavaleiro Rodrigues de Carvalho and submitted to the Mechanical Engineering
Department of the Faculty of Sciences and Technology of the University of Coimbra.
SURFACE MODIFICATION OF MOLDS AND ACESSORIES FOR THE GLASS INDUSTRY
Filipe Daniel Fernandes
IMAGEM
Bolsa de Doutoramento (SFRH/BD/68740/2010)
Acknowledgments
i
Acknowledgments
Many people have contributed and encouraged me along all the years of this research, I thank
all of them, and I hope I have gathered a bit of their best.
Above all, I want to express my sincere gratitude to my supervisors: Albano Cavaleiro and
Altino Loureiro, for their scientific guidance, commitment and availability during the
execution of this thesis. Their bright mind and wisdom brought to me many insights and
knowledge that was the key success for this thesis. I’m also extremely indebted to them for all
the opportunities I had of involvement with various enterprises and scientific research centers
as well as of visiting several wonderful places in the world in the framework of conferences
and collaborations.
I have had the pleasure of working with and learning from a huge number of other professors
and researchers in different countries. I would like to specially thank professors Amilcar
Ramalho, Tomas Polcar, Josep Guilemany and Jerzy Morgiel for all the collaboration,
scientific discussion and open access to their laboratories.
CEMUC in the person of its chairman, Professor Valdemar Fernandes, is greatly
acknowledged by the possibility of carrying out most of this work within this research center.
Thanks are also due to INTERMOLDE company for the support and collaboration in the
initial stage of the thesis. Thanks are also extended to TEandM Enterprise.
It is essential to emphasize the crucial role of IPN (Instituto Pedro Nunes) on my work and
results, by making available laboratorial equipment for the development and characterization
of the samples I produced.
Special thanks to my friends and research colleagues from ECAT, Surface Engineering and
Nanomaterials and Micromanufacturing groups of CEMUC. I have spent marvelous moments
in their companionship. These thanks are extended to all the young and senior members from
Acknowledgments
ii
the different international groups (Barcelona, Prague, Southampton and Guangxi (China)) I
visited, who welcomed me and helped me during my training periods there.
To my close friends thank you for all the great moments.
I would like to acknowledge the funding provided by the FCT (Fundação para a Ciência e
Tecnologia), through SFRH/BD/68740/2010 fellowship and the projects AUTOMATIN
“5380”, PLUNGETEC “13545” and PTDC/EME-TME/122116/2010.
Finally, I would like to thank my parents, grandparents and sister who permanently gave me
their support during this hard but reward time. Without their patience and good advices, I
would not have been able to take advantage of the opportunity I was given to improve my
education. This thesis is dedicated to them. Last but not the least, I thank you Cláudia for your
daily personal support and encouragement in all good and bad times.
Abstract
iii
Abstract
Coatings are frequently applied on molds and accessories for the glass industry in
order to restrict surface degradation such as oxidation, corrosion, abrasion and wear of the
structural material, thereby decreasing the maintenance costs and increasing the lifetime and
performance of the components. However, in order to obtain accurate lifetime expectancies
and performance of the coatings it is necessary to have a complete reliable understanding of
their properties.
This thesis is on the improvement of the surface properties and integrity of molds, in
order to increase their durability, through the application of different types of coatings. Two
methodologies were followed to reach such demands: (i) optimization of coatings currently
used in molds surface protection (Ni-based alloys deposited by Plasma Transferred Arc -
PTA); (ii) synthesis and characterization of new coatings with improved functionalities,
deposited by emergent deposition processes such as APS - Atmospheric Plasma Spraying
(effect of nanostructured ZrO2 additions on Ni-based alloy coatings) and DCRMS - Direct
Current Reactive Magnetron Sputtering (influence of V additions on the properties of
TiSi(V)N thin films).
The dilution of the substrate in PTA process was shown to strongly influence the
structure and, consequently, the hardness, the oxidation resistance and the tribological
behavior of the coatings; with increasing dilution, a detrimental effect on these properties was
observed due to the incorporation of base material. However, in relation to the tribological
behavior, a beneficial effect at high temperature was demonstrated due to the fast formation of
oxide layers which protect the coating surface against wear. The post-weld heat treatment
performed at coatings reduced the hardness of the partially melted and heat affected zones
without affecting the coatings hardness; whereas the coatings hardness and wear resistance
was improved with annealing treatment. Thus, the best performing coating could only be
achieved by a proper selection of the deposition conditions, in order to get the best
compromise between mechanical properties, high temperature oxidation behavior and wear
resistance of the coatings.
The impact promoted by nanostructured ZrO2 additions on the microstructure of a Ni-
based alloy depended on the way how the coatings were deposited by APS process, using: (i)
nanostructured ZrO2 and Ni-alloy powders previously mixed by mechanical alloying or (ii)
Abstract
iv
the same powders supplied separately. A homogeneous and compact microstructure with
small zirconia particles evenly distributed in the matrix was achieved in the first case, while a
porous microstructure, full of semi-melted Ni powders with large particles of ZrO2 entrapped
in their boundaries, suggesting a brittle behavior, was deposited in the second. In both cases
the hardness and wear behavior of ZrO2 rich coatings were improved in relation to the Ni-
based alloy. The coatings deposited from mechanically alloyed powders revealed to be much
more tribologically performing due to their compact structure and even distribution of
zirconia particles. All the APS coatings showed higher hardness values than the Ni-based
coatings deposited by PTA; however, their micro-abrasion resistance was worse, due to the
lack of cohesion between the powders.
The analysis by XRD of the structure of V rich TiSi(V)N coatings, deposited by
DCRMS, revealed that V incorporation in the TiSiN system shifted the peaks to higher
angles, indicating the formation of a substitutional solid solution based on TiN phase, where
Ti atoms are replaced by the smaller V ones. On the other hand, by similar reason, XRD of
TiSiN films revealed that a nanocomposite structure consisting of TiN grains enrobed by a Si-
N matrix was not formed. In fact, with Si addition a shift of the diffraction peaks of TiN phase
to higher angles was observed which, in combination with similar compressive residual
stresses, also supported the formation of a substitutional solid solution. V additions showed to
successfully improve the hardness and tribological behavior of TiSi(V)N films, as a result of
the substitutional solid solution formation. On the other hand, the formation of the V2O5 phase
during the sliding contact acts as a lubricious tribo-film, protecting the coating against wear.
The addition of Si or V strongly influenced in opposite directions the oxidation resistance of
the coatings. Si incorporations in TiN increased significantly the oxidation resistance of the
films, whereas the opposite occurred in V containing coatings. In this latter case, the rapid V
ions out-diffusion through the oxide scale inhibited the formation of a continuous protective
silicon oxide layer, which is responsible for the excellent oxidation behavior of TiSiN films.
V rich coatings showed lower oxidation resistance than PTA-deposited Ni-based coatings, but
superior hardness values and better tribological behavior was found in relation to both PTA
Effect of arc current on microstructure and wear characteristics of a Ni-based coatingdeposited by PTA on gray cast iron
F. Fernandes a,b,⁎, B. Lopes b, A. Cavaleiro a, A. Ramalho a, A. Loureiro a
a CEMUC, Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis Santos, 3030–788 Coimbra, Portugalb Intermolde, Rua de Leiria, 95, Apartado 103, 2431–902 Marinha Grande, Portugal
⁎ Corresponding author at: CEMUC, Department of Mecof Coimbra, Rua Luís Reis Santos, 3030–788 Coimbra, Porfax: +351 239 790 701.
Article history:Received 16 December 2010Accepted in revised form 3 March 2011
Keywords:Plasma transferred arcNi based alloyMicrostructureWear
The plasma transferred arc (PTA) technique is currently used to coat the edges of moulds for the glass industrywith nickel-based hardfacing alloys. However the hardness and wear performance of these coatings aresignificantly affected by the procedure adopted during the deposition of coatings. The aim of the presentinvestigation is to study the effect of arc current on the microstructure, hardness and wear performance of anickel-based hardfacing alloy deposited on gray cast iron, currently used in molds for the glass industry.Microstructure, hardness and wear assessments were used to characterize the coatings. Electron probe microanalysis (EPMA) mapping, scanning electron microscopy/energy dispersive X-ray analysis (SEM/EDAX) andX-ray diffraction (XRD) were used to characterize the microstructure of the deposits. The effect of post weldheat treatment (PWHT) on the microstructure and hardness was also studied. The typical microstructure ofthe coatings consists of dendrites of Ni–Fe, in the FCC solid solution phase, with interdendritic phases rich inCr–B, Ni–Si and Fe–Mo–C. Increasing the arc current reduces the proportion of porosity and hardness of thecoatings andmodifies their composition due to the increasing dilution of the cast iron. The partial melted zone(PMZ) had a typical white cast iron plus martensite microstructure, while the heat affected zone (HAZ) hadonly a martensite structure. The wear tests showed decreasing wear resistance with decreasing hardness ofthe coatings. PWHT reduces the hardness of the PMZ and HAZ but does not significantly alter the hardness ofthe bulk coating.
Traditionally, molds for the glass industry are manufactured fromgray cast iron or copper alloys due to their superior thermal behavior,relatively low cost and excellent thermal conductivity. This isassociated with high hardness and good wear resistance and leadsto a high production rate of glass parts. However, the edges of themolds made of these materials are sensitive to wear in heavy dutycycles; so they need to be coated with materials that are moreresistant to high temperature and wear [1].
The coating of these edges is frequently done by plasma transferredarc (PTA), using nickel-based fillermetals [2]. Normally the PTA processproduces very high quality, thicker deposits, offering optimal protectionwithminimal thermal distortion of the parts, low environmental impactand high deposition rates in single layer deposits [3]. It allows perfectlycontrolled deposition of alloys on mechanical parts that are subject toharsh environments, significantly extending their service life [4]. Thecomposition and properties of the coating are greatly influenced by the
type and dilution of the substrate. Balasubramanian et al. [5] suggestthat the dilution of coatings should be controlled through the relevantprocess parameters (transferred arc current, travel speed, powder feedrate, torchoscillation frequency and stand-off distance). Takano et al. [6]studied the influence of arc current variation, arc constriction andplasma gas flow on the characteristics of Co coatings. They concludedthat current intensity is the parameter thatmost significantly affects thecharacteristics of the deposits as higher currents increase the dilution ofthe base material and decrease the hardness of the coatings. Díaz et al.[7] mentioned that the dilution also increases with gas plasma flow instellite 6 coatings.
Nickel-based hardfacing alloys have become increasingly popularin recent years owing to their excellent performance in environmentswhere abrasion, corrosion and elevated temperature are factors. Themicrostructure of Ni-based hardfacing alloy deposits has been studiedusing various alloy compositions and different substrates [8–10].However, most of the studies consider the deposition of nickel alloyson stainless steels; there are no references to characterization of suchdeposits applied to gray cast iron.
Gurumoorthy et al. [8], who worked with nickel-base superalloysdeposited by PTA on stainless steel, reported that the microstructure ofthe deposit consists of γ-Ni solid solution dendrites, carbides, andinterdendritic eutectics composed ofγ-Ni and other phases identified as
being Ni-rich borides. The microstructure and mechanical properties ofNi base alloys are greatly influenced by the alloying elements. Forexample aluminum, titanium and niobium are added to strengthen thematerial through the formation of γ′ gamma prime (Ni3(Al,Ti)) andboron and zirconium are added to improve the creep strength andductility [9,10].
This paper details an investigation of themicrostructure, hardness and3-body abrasion behavior of a nickel-based hardfacing alloy deposited byPTA on gray cast iron with different weld currents. Microstructuralchanges in theHAZaswell as the effect of thepostweldheat treatment onthe microstructure and hardness of the coatings are also examined.
2. Experimental procedure
ThecoatingsweredepositedbyPTAtechnology (usingaCommersaldGroup ROBO 90machine) on flat surfaces of gray cast iron blocks using anickel-based hardfacing alloy. This technique obtains thicker coatingswith lower porosity and higher production rate than the Flame Sprayingand Gas Tungsten Arc Welding processes. The composition of base anddeposit materials is given in Table 1. The same process parameters wereused for eachdepositionwith theexceptionof the arc current,whichwas100 A for specimen 1, 128 A for specimen 2 and 140 A for specimen 4.The edges of a mold made of the samematerial but with a machined U-shaped groove were coated with the same alloy using the parameters ofspecimen 2. This specimen is referred to as specimen 3. The surfacingparameters used in the tests are listed in Table 2. The blocks of basematerial were induction heated before coating, in order to reducethermal shock, thus reducing the tendency to crack. After coating, theblocks were left to cool to room temperature in still air. Samplescontaining coating andbasematerialwere removed fromeachblock andthe mold for further analysis.
The samples for microstructural analysis were polished and etchedfollowing conventional procedures. Etchingwas performedusingequalparts of nitric acid and glacial acetic acid, to reveal themicrostructure ofthe deposits. Marble's solution (HCl (50 ml)+CuSO4 (10 g)+H20(50 ml)) was used to reveal the microstructure of the cast iron. Themicrostructure of the different regions of the samples was characterizedby optical microscopy (OP), scanning electron microscopy (SEM), X-raydiffraction (XRD) and energy dispersive X-ray analysis (EDAX). Thechemical composition was evaluated by Electron Probe Microanalysis(EPMA). The hardness profile across the interface of the coatings wasdetermined with Vickers testing using a 5 N load at 15 s.
Some specimens were subjected to a subsequent heat-treatmentat 850 °C for 1 h and then cooled in the furnace to room temperature,in order to reduce brittle phases present in the partially melted andheat affected zones. Microstructural and hardness profiles of thesespecimens were obtained.
Wear tests on the coated surfaces were carried out at roomtemperature in ball-cratering devices. In these devices, a ball is placedon a flat specimen in the presence of abrasive slurry and rotated whilebeing submitted to a specific load in order to produce a spherical cup-shaped depression. This process was used in order to reproduce thewear mechanism present at the surface of the coatings on glass moldswhich have been in service for a long time. An AISI 52100 steel ballbearing 25.4 mm in diameter was used at constant rotational speed of75 rpm. High purity SS40 silica with angular particles averaging
Table 1Nominal chemical composition (wt.%) of substrate and hardfacing alloy.
Base material C Mn Si P S
Grey cast iron 3.60 0.60 2.00 b0.20 b0.04
Hardfacing alloy C Cr Si
Ni-alloy 0.14 2.45 2.56
3.10 μm in size was used as an abrasive to replicate the effect of themelted glass. The slurry used was prepared with a proportion of28 vol.% of silica in water.
The duration of the test, measured by the number of rotations, wasselected according to the material and the test conditions employed.The normal load used was 0.1 N.
Photographs of the spherical depressions were taken in order tomeasure their dimensions. The wear volumewas calculated by Eq (1),where R is the ball radius and b is the crater chordal diameter.
V = π × b4� �
= 64 × Rð Þ ð1Þ
Archard's model, given by Eq (2), is the onemost commonly appliedto wear results to determine a parameter to quantify the wear behaviorofmaterials. In this equation P is the normal load, l is the sliding distanceand k is the specific wear rate. The specific wear rate is the parameterwhich quantifies the wear behavior.
V = k × P × l ð2Þ
To quantify the specific wear rate a linear version of Archard'smodel was applied and statistical analysis was used to estimate theerror of measurement [11].
3. Results and discussion
3.1. Microstructure
Four distinct regionswere observed in the cross section of each of thecoatingsdepositedwith thedifferentwelding conditions, as illustrated inFig. 1 A) for sample 3. Those regions are the fusion zone (FZ), composedof the Ni-alloy and the base material melted during the process, thepartially melted zone (PMZ), which is the area immediately outside theFZ where liquation can occur during welding, the heat-affected zone(HAZ), which was not melted but underwent microstructural changesand the base material (BM), whose microstructure remains unaffectedduring the coating deposition thermal cycle. The same regions wereobserved by M. Pouranvari [12], who studied the welding of gray castiron using nickel filler metal deposited by shielded metal arc welding.
Cr Ni Mo V Ti Fe
b0.20 b0.50 0.50 0.10 0.20 Balance
B Fe Al F Co Ni
0.86 1.08 1.30 0.01 0.08 Balance
Fig. 1. (A) microstructure of various regions in gray cast iron weld in as-weld condition, (B) partially melted zone, (C) heat affected zone, (D) transition heat affected zone–basematerial, (E) base material.
Fig. 2. Dilution induced by PTA process using different weldments: specimen 1 — 100 A, specimens 2 and 3 — 128A, specimen 4 — 140 A.
4096 F. Fernandes et al. / Surface & Coatings Technology 205 (2011) 4094–4106
Fig. 1 B) to E) illustrates the microstructures of the zones marked withthe digits 1 to 4 in Fig. 1 A) in higher magnification.
Fig. 1 B) shows the transition zone between the coating and thepartially melted zone. Fig. 1 C) illustrates that the HAZ is basicallycomposed of a martensitic structure (needle shape), which is foundmainly in the region closer to the fusion line, however, precipitates andgraphite flakes are also present. Themicrostructure of the basematerial isprincipally ferritic with flakes of graphite and precipitates uniformlydistributed in the matrix (Fig. 1 E)). Fig. 1 D) illustrates the transitionbetween the HAZ and the base material. The nature and relative size ofthose zones are determined by the coating procedure, mainly the heatinput during the process, and the composition of the cast iron andhardfacing alloyused. The constitutionof themain zoneswill be discussedseparately, in order to provide a better understanding of the effect ofcurrent on their characteristics.
3.1.1. Melted zoneFig. 2 illustrates the effect of the current on the dilution produced
in each sample. The dilution is defined as the proportion between thearea of melted base material and the total area of the melted zone,
Fig. 3. Aspect of the coatings of: A) specimen 1, B) specimen
both measured from the cross section of the coating using Photoshopimage analysis. These images show that as the current is increased inthe range 100–140 A, the dilution steadily increases from 28% to 59%.Specimen 3 in the image illustrates the groove shape usually machinedin the mold before coating. Comparing specimens 2 and 3, which wereboth coated using the same current, it is possible to see that the grooveproduces only a small change in the dilution and the results obtainedfrom the flat blocks can be extrapolated to molds.
The dilution of cast iron modifies the composition of the coatings,as illustrated in Table 3, which shows the chemical compositionmeasured by EPMA at various points of the coating and the substrateof specimen 3. Point 1 is located close to the coating surface, point 2 atthe mid-thickness of the coating, point 3 in the coating close to theinterface and the other two points are in the bulk cast iron; point 4close to the interface and point 5 distant from it. The amount of iron inthe coating (N16%) is much higher than that provided by thehardfacing alloy (1.08%), producing a consequent reduction in thenickel content. According to Ezugwu et al. [9] the increase in ironcontent in the coatings tends to decrease their oxidation resistancebecause of the formation of a less adherent oxide scale. The increase inMo content in the coating is also caused by the dilution of the cast ironbecause it is absent in the hardfacing alloy. Although the true carboncontent value could be skewed by contamination, carbon is present inthe coating in large quantities, again due to the contribution of thegraphite from the cast iron. Thus, increasing the arc current willincrease the dilution and the chemical composition of the coating willchange accordingly.
All coatings contained small random pores throughout the crosssection, as shown in Fig. 3. According to F. A. L. Dullien [13], the“volumetric” and “area” porosity proportions are equivalent for random-structured porous media, which agrees with our results. The area of the
2, C) specimen 4, and D) fusion interface of specimen 1.
Fig. 4. Optical micrograph of: A) specimen 1, B) specimen 2, C) specimen 4, and D) interface specimen 1.
4098 F. Fernandes et al. / Surface & Coatings Technology 205 (2011) 4094–4106
poresvisibleunder theopticalmicroscopewascalculatedusing the ImageJprogram (developed at the National Institutes of Health, United States).The proportion of porosity calculated is very low for all specimens; 0.35%for specimen 1 with an average pore size of 1.45±0.58 μm, 0.32% forspecimen 2 with an average pore size of 1.61±0.65 μm and 0.28% forspecimen 4 with an average pore size of 1.37±0.40 μm. This decrease inthe proportion of micropores in the coatings as current increases can becorrelatedwith the refinement of thedendriticmicrostructurementionedbelow. The decrease in dendrite size reduces the amount of liquid metaltrapped by dendrites during solidification, reducing the number of voidsin the microstructure. Charmeux el al [14] mentioned this mechanism inthe solidification of Almicro castings.On the other hand the cast iron alsoexhibits significant porosity, with pores reaching 17 μm in diameter, asillustrated in Fig. 3 D).
The typical microstructure of the coatings is shown in Fig. 4. Themicrostructures consist of dendrites of the Ni–Fe solid solution phase,with columnar morphology oriented along the direction of heat flowand torturous grain boundaries. The microstructure contains C flakes(dark floret-like structures) equally distributed through the micro-structure. The chemical composition of the Ni-based powder (lowcarbon content see Table 1) does not explain the large number ofthese flakes which can only be attributed to the dilution of cast ironinduced by the PTA process. Furthermore, small interdendriticprecipitates can also be detected. These are possibly carbides, asexplained below. Fig. 4 A–C) shows that an increase in the arc currentand, therefore, an increase in dilution, allows carbides and C-flakes inthe microstructure to be more easily detected and observed. It wasthought that there could be some coarsening of the microstructure as
the current was increased due to greater heat input. However this wasnot observed. Fig. 4 shows that the dendritic structure becomes fineras the current increases. For example, the average ternary dendritespacing for the specimens 1 and 4 is 15.4 and 11.2 μm, respectively. Asexplained below, this refinement can be related to the change in thecomposition of the deposited material due to its increasing dilution asthe arc current increases.
The phenomenon of circular “islands”, which frequently occur close tothe interface of the coating, is common to all the specimens (see Fig. 4 D)for specimen 1), These islands are characterized by the total absence ofgraphite. According to the Commersald company [15] they are causedwhen powder grains not melted by the plasma arc subsequently melt incontact with the pool. Near the interface a cellular microstructure finerthan thedendritic structurewasobserved for all the samples, as illustratedin Fig. 4 D) for specimen 1, where the grain size is approximately 12 μm.This refinement may be due to the higher solidification rates involved inthe boundary because of the efficient thermal exchange ensured by thehigh volume ratio substrate/coating of the base material, as suggested byGatto et al. [3].
SEM analysis gives amore detailed view of themicrostructure of thecoatings. Fig. 5 shows a SEM image of the microstructure of the coatingof specimen 1. The image shows that in addition to the dendrites of theNi Fe solid solution phase and graphite flakes the grain boundary isformed of two phases: a light gray phase (like-grain) and a dark grayone. EDAX spectra showed that the light gray phase and the dark grayphase are rich in silicon and chromium, respectively.
A detailed analysis was carried out to study themain constituents ofthe coatings using EPMA. WDS (wavelength dispersion spectroscopy)
Fig. 5. A) Typical SEM micrograph (etched) revealing the grain boundaries of specimen 1; SEM EDAX spectra of: B) dark gray phase, and C) light gray phase.
mapsof themost abundant elements of the coating for all the specimenswere obtained. Fig. 6 A) represents a SEM image of themicrostructure ofthe deposit for the specimen 1. The respective elemental maps of Fe, Ni,Si, Cr, Al, B, C andMo are shown in Fig. 6 B–I), color-coded such that thelowest concentration of the element analyzed is indicated in purple andthe highest in red. The results of the WDS maps for specimen 1 revealthat the dendrites are rich in iron, silicon and aluminum, whileboundaries are rich in boron, silicon and chromium.Nickel is distributedalmost uniformly throughout the structure, except in areas wheregraphite flakes are present. In the boundary chromium appears, toexclude siliconas these elements do not overlap. However, chromiumand boron do overlap in the boundaries, as illustrated in Fig. 6 E) andH).Thus, it can be concluded that the borders of the boundary are rich in Siwhereas themiddlesof theboundaries aremainly composedof B andCr.Comparing the information from Figs. 5 and 6 it is possible to conclude
that the light gray (like-grain) phase is rich in silicon and the dark grayphase is rich in chromium,which agreeswith the EDAX analysis. Carbonconcentrates chiefly in the graphite flakes, as shown in Fig. 6 F), thoughat someparticular points the increment in theC signalwasalsoobservedin the boundary, mainly coincidingwith the B signal (see circled zone inB andCmaps) suggesting that carbidesmayhaveprecipitated. Althoughnot visible in the C-map, very small agglomerates ofMo can be detectedin Fig. 6 also suggesting formationof carbides.However, as shownbelowby XRD analysis, these carbides were not identified, probably due totheir small size and/or content.
The increase in the arc current and, consequently, in thedilutionof castiron brought some changes in the distribution of themain elements in themicrostructure, as shown by theWDSmaps for specimen 4, illustrated inFig. 7. For this specimen, in comparison to sample 1, themain difference isrelated to the Ni and Fe signals. It is clear from Fig. 7 that the zones in the
Fig. 6. WDS maps of the coating of specimen 1: (A) SEM image of the microstructure of the deposit, (B) iron, (C) nickel, (D) silicon, (E) chromium, (F) carbon, (G) aluminum, (H)boron, and (I) molybdenum.
4100 F. Fernandes et al. / Surface & Coatings Technology 205 (2011) 4094–4106
interdendritic boundaries richer in iron are depleted in nickel and siliconbut that in the boundary these zones overlapwith Cr and B. The groupingof thesemaps suggests the presence of a Fe–Cr–Bphase in the boundaries,however, as can be seen later from the XRD analysis this is not identified,as only a preferential associationwith the B–Cr phase is indicated, leadingfree iron to combine with other elements. The Si rich zones, see Fig. 7 D),are depleted of Fe and rich inNi. In specimen4, aswell as the strong signalgiven by the C flakes tiny molybdenum and carbon traces could bedetected in the surroundingmaterial (whereC is hardlypresent),which isa clear contrast with the results from specimen 1. These elements aredistributed preferentially in the boundaries overlapping the Fe map,suggesting the formation of a mixed carbide containing Fe and Mo. Theincrease in the content of these elements is caused by the increased
dilution, since they are present only in the cast iron. Collins and Lippold[16] and Ramirez et al. [17] mention that the presence of precipitates inthe interdendritic regions results in the formation of very tortuous grainboundaries and theexcess of theses precipitates stops themigrationof thegrain boundaries. This can justify the lower grain size of the coatingdepositedwith 140 A in spite of the higher energy supplied to the coatingprocess. The literature also indicates that an increase in the molybdenumand carbon content improves the rate of heterogeneous transformationduring solidification. These elements normally segregate in the grainboundaries, thereby hindering grain growth [18]. Nevertheless, theinfluence of the much higher number and content of graphite flakespresented in the 140 A sample, these being segregated from the grainboundaries during material solidification, should therefore also be
Fig. 7. WDS maps of the coating of specimen 4: (A) SEM image of the microstructure of the deposit, (B) iron, (C) nickel, (D) silicon, (E) chromium, (F) carbon, (G) boron, and (H)molybdenum.
considered as anexplanation for the lowergrain size in this sample. As canbe seen in Fig. 4, C flakes are common in the grain boundaries.
The distribution of the chemical elements in the microstructure inspecimens 2 and 3 follows the trend mentioned above.
Fig. 8 shows the XRD spectra of specimens 1, 2 and 4. The analysisof this data revealed that the major phase present in the coatings is a(Ni, Fe) solid solution face centered cubic (ICDD card 47–1405-(111),(200), (220), (311), and (222) peaks at 2θ~51.1°, 59.8°, 89.6°, 111.42°and 119.3°, respectively). WDSmaps show that in Cr and B rich zones,Si does not exist as it is connected to Ni. The overlapping of Ni and Sisignals in WDS maps suggests the presence of a Ni–Si phase, such asNi3Si (ICDD card 32–0699) shown in Fig. 8, corresponding to some of
the lower intensity experimental XRD peaks. The other peaks can beadjusted either to a Cr-boride phase Cr5B3 (ICDD card 32–0278) or tothe Fe, Mo mixed carbide of M6C (Fe3Mo3C — ICDD 47–1191) type.The increasing content of this latter phase through samples 1 to 4 sitswell with the WDS maps presented and interpreted above. Further-more, the indexation is in agreement with the results from theliterature where similar phases were also detected in a nickel alloydeposit on an austenitic stainless steel [8]. In summary, the increase inthe arc current increased the dilution of the cast iron, raising the iron,carbon and molybdenum content in the coating, promoting dendriterefinement of the microstructure due to the increase in precipitates(Fe–Mo–C) and C flakes in the grain boundaries.
Fig. 8. X-ray diffraction pattern of the deposit material obtained for the specimens 1, 2 and 4.
4102 F. Fernandes et al. / Surface & Coatings Technology 205 (2011) 4094–4106
3.1.2. Partially melted zone (PMZ) and heat affected zone (HAZ)These zones are critical in cast irons since the material can solidify
as white iron if cooling is rapid enough. If the amount of graphitedissolved during welding is high enough it is likely that it will alsogive rise to a continuous carbide network. This is undesirable as abrittle carbide matrix can cause durability problems in the final part.Fig. 9 shows the microstructure of this zone in specimen 3. In thisimage, cementite (Fe3C—white phase) can be seen concentrated nearthe interface. Fig. 9 B) shows the microstructure of the PMZ zoneunder high magnification. The image displays a large amount ofcementite in the grain boundaries with acicular martensite inside thegrains. Some precipitates, identified as mixed titanium and molybde-nium carbides (marked respectively with the numbers 1 and 2 inFig. 9 and identified in Fig. 10, were also found. As these phases arehard and brittle this region is sensitive to in-service crack initiationand propagation due to thermal and mechanical fatigue, as discussedbelow. M. Pouranvari [12] also reported that a continuous brittlenetwork of coarse carbides along the weld fusion line may lead toinitiation of cracking.
In the heat affected zone, carbon can diffuse into the austeniteduring welding and the austenite may subsequently transform intobrittle martensite due to the high cooling rate. Martensite is also
Fig. 9. SEM image of partial melted zone and heat affected zone of specimen 3, (A
susceptible to cracking. The amount of martensite formed depends onthe composition of the cast iron and the thermal history of the zone.
3.2. Hardness
Fig. 11 shows the hardness profiles across the interface for all thestudied specimens. The vertical line in the graph represents the fusionline. The hardness through the melted material is approximatelyconstant and tends to decrease with increasing weld current. This canbe related to the higher dilution induced by the process.
For all the samples, the highest hardness valuesweremeasured in thearea near the fusion line. A maximum hardness of 538 HV was observedin specimen 1 and decreases with increasing arc current for the otherspecimens. As previously documented, themicrostructure in that regionconsistedofhardmartensite andcementite. Theproportionofmartensitein themicrostructure decreases with increasing distance from the fusionline and, thus, the hardness decreases too. Themicrostructure in the HAZis directly related to the heat input in the process and therefore to thecurrent intensity used.
The high hardness values found in the PMZ and HAZ make theseregions potentially responsible for many of the mechanical problemsexperienced in welds in cast iron. The most effective way to reduce
) transition between coating and the substrate, and (B) magnification of PMZ.
Fig. 10. Energy dispersive X-ray analysis (EDAX) of the particles shown in Fig. 9 b), (1) titanium carbide with square shape, and (2) molybdenum carbide with rounded shape.
Fig. 11. Hardness profile across the interface.
Fig. 12. Hardness profile across the interface after PWHT in specimen 4.
4104 F. Fernandes et al. / Surface & Coatings Technology 205 (2011) 4094–4106
hardness and cracking problems in these regions is to reduce theseverity of the thermal cycle produced by the process. This can be doneby controlling the heat input and preheating during coating. Anotherway to reduce the hardness in these zones is to submit the parts to anannealing treatment, although this solution is expensive.
The precipitates, the porosity and the random distribution of graphiteflakes in the cast ironmatrix can be linked to the variability of hardness inthe base material.
Fig. 14. Evolution of the wear behavior as a function of the parameter “normalload×sliding distance”.
3.3. Effect of post-weld heat treatment (PWHT) on the microstructureand hardness
Although the heat treatment did not induce significant changes inthe microstructure and hardness of the coatings, it induced substan-tial changes in the PMZ and HAZ, as illustrated by the hardnessprofiles before and after PWHT shown in Fig. 12 for specimen 4. Thisspecimen was produced with the highest arc current, and had thelowest hardness values in the PMZ, as shown in Fig. 11. Fig. 12 showsthat even in this case the heat treatment (holding at 850 °C for 1 h)produced little change in the coating but a substantial decrease in thehardness of the PMZ and HAZ of the specimen. Similar behavior wasobserved for the other specimens. This reduction in hardness can beattributed to the reduction in the proportion of hard phases,cementite and martensite, in these zones. After PWHT the basematerial displays basically a perlitic/ferritic structure.
Fig. 13. Indentations in base material (specimen 4) after PWHT. A) indentation made inthe perlitic structure, and B) indentation made in the ferritic structure.
After heat treatment a large scatter of hardness measurementscontinues to be observed in the cast iron. This scatter can be attributedto local alterations in the microstructure, as shown in Fig. 13, whichrepresents the aspect of the indentations marked with the numbers 1and 2 in the hardness profile in Fig. 12. The image reveals that if theindentation impacts mainly the perlitic structure, the white zone inFig. 13 A), the hardness will be greater than if the indentation isperformed in the ferritic structure, the gray zone in Fig. 13 B). Inconclusion, PWHT is beneficial because it reduces hardness in the PMZandHAZ,without producing significant alterations in themelted zone.
3.4. Wear test
Thewear tests were performed to predict the effect of the differentprocess parameters on the in-service wear behavior of the coatedcomponents. The ball-cratering test results are shown in Fig. 14, as thevolume of material loss plotted against the product of sliding distance(l) and normal load (P). A complete reliability analysis is summarizedin Table 4. The wear volume displays linear evolution with P×l, andthe slope corresponds to the specific wear rate. Specimen 1 exhibitsboth the lowest wear volumes and the lowest specific wear rate,therefore, it is the most resistant to wear. This is compatible with thegreatest hardness displayed by the coating of this specimen, asillustrated in Fig. 11. Samples 2 and 4 display very similar wearvolumes, which is confirmed by similar hardness values; even so,sample 2 has a specific wear rate slightly higher than sample 4.
Fig. 15 A) shows a SEM micrograph of a wear scar induced by theball-cratering device and Fig. 15 B) the surface of a coating of a moldafter a long time in service, which allows pitting to be identified as themajor failure mechanism. The image also shows some cracks that canbe attributed to the localized melting of hard, brittle phases.
Table 4Results of the linearization analysis.
Specific wear rate,k (mm3/N m)
Average STD Confidenceinterval 90%
r2
Specimen 1 6.9×10−4 5.2×10−5 5.7×10−4 to 8.0×10−4 0.977Specimen 2 9.6×10−4 6.6×10−5 8.2×10−4 to 1.1×10−3 0.982Specimen 4 8.0×10−4 4.6×10−5 7.0×10−4 to 9.0×10−4 0.987
Ball cratering is a micro-abrasion technique in which testconditions can be tailored to adjust the wear mechanism to grooving,which is plastic deformation controlled, or rolling, which is a fracturecontrolled process [19]. Prior tests were used to select test parameterscompatible with rolling wear, designed to induce pitting failure.
The wear scar displays two modes of wear: pitting and grooving,pitting being the more extensive. Grooving occurs when particles sliponto the surface. Fig. 16 shows the morphology of these two wearmodes in higher magnification. The wear mechanism present in thespherical cup depressions (essentially 3-body rolling wear) accuratelyrepresents the wear mechanism in the molds. However one shouldkeep in mind that the molds in service are subject to other damagemechanisms such as thermal and mechanical fatigue and corrosion athigh temperature.
4. Conclusions
The influence of the arc current used in the plasma transferred arcprocess (PTA), as well as the effect of post-weld heat treatment, on themicrostructure, hardness and wear resistance of deposited layers wasanalyzed in this study. The increase in the arc current increased thedilution of the cast iron, changing the composition andmicrostructureof the coatings. Moreover, the increase in current raises the quantityof precipitates and C flakes at the grain boundaries, which leads to therefinement of the microstructure. Increasing the arc current gives riseto a reduction in the hardness in all regions of the coatings. The
reduction in the hardness of the coatings is accompanied by areduction in their wear resistance in wear tests performed at roomtemperature with silica slurry. Post-weld heat treatment reduces thehardness in the partially melted zone and heat affected zone asopposed to the coatings themselves, where hardness remainsunaffected.
In the future, abrasion and oxidation tests at high temperature willbe done, in order to reproduce the conditions of industry application.
Acknowledgements
The authors wish to express their sincere thanks to the companyIntermolde, to Instituto Pedro Nunes (IPN) and the PortugueseFoundation for the Science and Technology (FCT) through COMPETEprogram from QREN and to FEDER for financial support.
References
[1] M. Cingi, F. Arisoy, G. Basman, K. Sesen, Mater. Lett. 55 (2002) 360.[2] E.B. Hinshaw, ASM HANDBOOK, 1993, p. 740.[3] A. Gatto, E. Bassoli, M. Fornari, Surf. Coat. Technol. 187 (2004) 265.[4] K. Siva, N. Murugan, R. Logesh, Int. J. Adv. Manuf. Technol. 41 (2009) 24.[5] V. Balasubramanian, A.K. Lakshminarayanan, R. Varahamoorthy, S. Babu, J. Iron.
Steel Res. Int. 16 (2009) 44.[6] E.H. Takano, D. Queiroz, A.S.C.M. D' Oliveira, Soldagem Insp. São Paulo 13 (2008)
210.[7] V.V. Díaz, J.C. Dutra, A.S.C. D' Oliveira, Soldagem Insp. São Paulo 15 (2010) 41.[8] K. Gurumoorthy, M. Kamaraj, K.P. Rao, A.S. Rao, S. Venugopal, Mater. Sci. Eng., A
F. Fernandes, A. Cavaleiro, A. Loureiro, Oxidation behavior of Ni-based coatings
deposited by PTA on grey cast iron, Surface and Coatings Technology, 207 (2012)
196-203.
Surface & Coatings Technology 207 (2012) 196–203
Contents lists available at SciVerse ScienceDirect
Surface & Coatings Technology
j ourna l homepage: www.e lsev ie r .com/ locate /sur fcoat
Oxidation behavior of Ni-based coatings deposited by PTA on gray cast iron
F. Fernandes ⁎, A. Cavaleiro, A. LoureiroCEMUC - Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis Santos, 3030‐788 Coimbra, Portugal
Article history:Received 28 February 2012Accepted in revised form 20 June 2012Available online 29 June 2012
Keywords:High temperature oxidationOxide scalesSurface morphologyNickel alloysPlasma transferred arcThermogravimetric measurements
The aim of this investigation was to study the effect of PTA current (100 and 128 A) on the oxidation behaviorof nickel-based hardfacing coatings deposited on gray cast iron. The oxidation behavior of coatings held at800 and 900 °C for 2 h in air was studied by thermogravimetry (TGA). The surface, aswell as the cross‐section,of the coatingswas characterized by scanning electronmicroscopy combinedwith energy-dispersive X-ray spec-troscopy (SEM/EDS) and X-ray diffraction (XRD). TGA results indicate that the coating produced with lower arccurrent exhibits more effective oxidation resistance than that producedwith higher current. This behavior couldbe correlated with the dilution promoted by the PTA process, which changes the chemical composition of coat-ings. As a consequence, different kinds of oxide scaleswere detected in each coating after isothermal oxidation. Inthe specimen producedwith a lower arc current and lower dilution a protective layer of Si–O is formed, while inthe specimen producedwith a higher current two layers could be identified: an external one of Fe2O3, with smallfeatures of Fe3O4 and NiFe2O4 spinel rich phases, and an internal one of Fe3O4with small amount of a dark phaseevenly distributed rich in silicon, nickel and iron. The isothermal oxidation curve at 900 °C of the coating withhigher dilution showed two stages: at an early stage, the weight increase over time is almost linear whereas,in a second stage, a parabolic law could be fitted to the experimental data. The other specimens followed onlya parabolic law.
Cast irons and copper-alloys are commonly used in the production ofglassmolds and accessories for the glass industry [1]. In service,molds aresubjected to very severe abrasion, wear, oxidation, and fatigue at hightemperatures, due to contact with melted glass, which limit their surfacelife. As a rule, automated manufacturing processes are operated withmold temperatures close to the critical temperature of themoldmaterial.Although the temperature of melted glass is about 1050 °C the surfacetemperature of molds is currently in the range of 500–900 °C, due to aninternal cooling system in the molds. Above a critical temperature theglass body starts to adhere to the mold surface. Higher temperatureslead to mold failure. Therefore, protective coatings are normally appliedto give themolds high temperature resistance, with the aim of increasingtheir lifetime in harsh environments [2,3].
Nickel-based superalloys gain an extremely important role in protec-tion of surfaces in components for the glass industry due to their excel-lent performance under conditions of abrasion and corrosion at elevatedtemperatures [4–6]. Several processes have been used to apply coatingsto protect the surfaces of these components, as follows: (i) thermalspraying processes such as high velocity oxy fuel (HVOF), atmosphericplasma spraying (APS) and flame spraying (FS); and (ii) hardfacing pro-cesses such as plasma transferred arc (PTA) and gas tungsten arc
: +351 239 790 701.ndes).
rights reserved.
welding (GTAW) processes. However, since hardfacing processes giverise to a metallurgical bonding with the substrate they have been usedpreferentially in the protection of mold surfaces as opposed to thermalspraying processes which produce only mechanical bonding of thelayer to the substrate. Among the hardfacing processes above, PTA hasbeen widely used to protect the surface of molds, especially theiredges, because they are submitted to heavy wear duty cycles. This pro-cess gives rise to thicker deposits of very high quality, providing a highdeposition rate in a single layer [7–9]. However, the chemical composi-tion and properties of the coatings, as well their quality, are stronglyinfluenced by the dilution of the substrate promoted by the PTAprocess.Low dilution provides coatings with a chemical composition similar tothe added metal powder, which is a prerequisite to getting improvedwear and oxidation resistance [10]. However, as dilution is decreasedby reducing process current the occurrence of defects such as lack of fu-sion increases,which is catastrophic in terms of the behavior ofmolds inservice. So during coating deposition, the tendency is to increase the di-lution of the base material even if oxidation resistance can be under-mined. As oxidation is one of the major failure factors at hightemperature in molds for the glass industry, it is imperative to under-stand how the different coatings provided by different process condi-tions behave under the harsh and oxidizing atmospheres.
Technical literature reports a lot of studies on the high temperatureoxidation behavior of nickel alloys. Liu et al. [11], who studied the oxi-dation behavior in air of a single-crystal Ni-base superalloy at 900 and1000 °C, reported two oxidation steps for both temperatures. The first
one is controlled by NiO growth and the second by Al2O3 growth until acontinuous Al2O3 layer formed under the previously grownNiO layer. Liet al. [12] studied the oxidation of a NiCrAlYSi overlayerwith orwithouta diffusion barrier deposited by one-step arc ion plating. They showedthat the duplex coating system exhibits a more effective protection forthe substrate, where thin and continuous scales are adhered to theoverlayer surface, and very limited oxidation and interdiffusion attacksare detected. Zhou et al. [13] studied the oxidation behavior of pure anddoped nickel alloys with different Co contents in air at 960 °C. They ob-served that increasing the Co content increases themass oxidation gainof the coatings.
Since oxidation is one of the main drawbacks that limit the life ofglass molds, the aim of this research is to study the effect of PTA currentvariation on the oxidation behavior of coatings deposited on gray castiron using a nickel-based alloy. The oxidation behavior of the coatingswas studied by thermal gravimetric analysis (TGA). The surface andcross‐section morphologies of coatings after isothermal oxidation wereobserved and characterized by scanning electron microscopy providedwith energy dispersion spectrometry (SEM-EDS) and x-ray diffraction.Further, the high temperature oxidation mechanisms are discussed.
2. Experimental procedures
Nickel based Colmonoy 215 (fromColmonoy Company) powderwasdeposited by plasma transferred arc (PTA) onto specimen blocks of graycast iron, currently used in the production of molds for the glass indus-try. The deposits were executed using a Commersald Group ROBO 90machine using two different arc currents (100 and 128 amperes). Thegoal of using two different arc currents was to achieve coatingswith dis-similar levels of dilution with the substrate. The nominal compositionsof the cast iron and Colmonoy 215 powder are displayed in Table 1.The principal surfacing parameters employed are shown in Table 2. Be-fore coating, the blocks of cast iron were induction heated at 480 °C, inorder to reduce susceptibility to cracking during coating deposition.
After coating, specimens containing cast iron and coating were re-moved from each block for microstructural analysis. Metallographicanalysiswas done using conventional procedures. Small samples for ox-idation testswere also removed from each coating. The surfaces of thesesamples were ground using number 1000 SiC abrasive paper in order toproduce even surface preparations. Further, the specimens were ultra-sonically cleaned in alcohol.
Isothermal oxidation tests were conducted at 800 and 900 °C in airfor 2 h in a thermal gravimetric analysis (TGA) machine. The air fluxused was 50 ml/min and the heating rate up to the isothermal temper-ature was 20 °C/min. After thermal exposure the samples were coolednaturally in air to room temperature. The weight gain of samples wasevaluated at regular 2 s intervals using amicrobalance with an accuracyof 0.01 mg. After oxidation the surface morphology of specimens was
Table 2Main deposition parameters for PTA weld surfacing.
Main arccurrent (A)
Powder feedrate (rpm)
Travel speed(mm/s)
Powder feed gas flowrate (l/min)
Plarat
Coating 1 100 20 2 2 2.2Coating 2 128 20 2 2 2.2
examined by scanning electron microscopy with x-ray spectroscopy(SEM-EDS). Further, the same equipmentwas used to identify the layersforming the oxide scale from the cross‐section of samples. In this casethe specimensweremounted in epoxy resin, polished and then surfacedwith a thin layer of gold for perfect observation. X-ray diffraction (XRD)using Co Kα radiation was conducted on the oxidized surface of speci-mens. In order to ensure the reproducibility of results, three specimenswere analyzed for each coating and set of test conditions.
3. Results
3.1. Microstructure of the as-deposited coatings
Fig. 1 displays the microstructure of the coatings deposited on thesurface of gray cast iron coupons. The coatings show dense micro-structure without lack of fusion of the substrate, free of microcracksand few solidification voids. Their typical microstructure consists ofdendrites of the Ni-Fe solid solution phase aligned along the directionof heat flow. Furthermore, it displays C-flakes (dark-floret like struc-tures) randomly distributed in the matrix. The increase in arc currentgives rise to a higher dilution of the gray cast iron, changing the orig-inal chemical composition of the filler metal. Following the proceduredetailed in reference [9], dilutions of 28% and 54% were measured forthe coatings deposited using 100 and 128 A arc current, respectively.These levels of dilution can explain the increase in C-flakes with in-creasing arc current, as shown in Fig. 1. Table 3 shows the chemicalcomposition measured by energy dispersive spectrometry (SEM-EDS)from an area of 400×400 μm, from the cross‐section at the middlethickness of each coating. With this procedure it is assured that chem-ical composition values are measured from a representative volume ofmaterial in that zone, avoiding erroneous measurements that couldarise from point analyses of the heterogeneous microstructure. As canbe seen, the main differences in chemical composition are related tothe higher iron, carbon and silicon contents achieved with increasingdilution. The detailed characterization of the microstructure of eachcoating can be found in a previous publication by the authors [9].
3.2. Isothermal oxidation in air
The results of the thermo gravimetric analysis performed on the coat-ings at different isothermal temperatures (800 and 900 °C) are shown inFig. 2a and b for the coatings deposited using 100 and 128 A arc currents,respectively. From now on and throughout the text, the coatings depos-ited using 100 and 128 amperes will be identified as “C100” and “C128”,respectively. The C100 coating exhibits parabolic oxidation weight gainas a function of time for both testing temperatures as does the C128 coat-ing at 800 °C. On the other hand, oxidation of the C128 coating at 900 °Cstarts with a linear increase in mass gain but after 1400 s it starts to
sma gas flowe (l/min)
Shielding gas flowrate (l/min)
Torch workdistance (mm)
Oscillation(mm)
Preheattemperature,(°C)
20 13 4 48020 13 4 480
Fig. 1. Typical microstructure of: A) coating deposited using 100 A of arc current, B) coatingdeposited using 128 A of arc current.
198 F. Fernandes et al. / Surface & Coatings Technology 207 (2012) 196–203
follow a parabolic path. The figures also reveals that for both oxidationtemperatures, coating C100 displays less oxidation weight gain thancoating C128. The dissimilar levels of oxidation observed in the speci-mens can be correlated with the dilution promoted by the PTA process,that changed their chemical composition, as shown in Table 3, whichwill interfere with the oxidation process. As mentioned before, themain differences in chemical composition of coatings are related to sili-con, carbon and, particularly, iron content. Wallwork [14] and Ezugwu[15] reported that if iron content is increased in nickel-based alloys, ittends to decrease their oxidation resistance, as is the case in our study.This behavior was attributed to a progressively higher cation diffusionrate in the scale. Therefore, increasing the dilution of gray cast iron by in-creasing the arc current proves to be detrimental, since the oxidation re-sistance of coatings is reduced due to a further increase of iron content inthe coatings. Moreover, it is observed that the oxidation gain of speci-mens increases with increasing isothermal temperature. This is anexpected effect since all diffusion phenomena ruling the oxidation pro-cess are enhanced as the temperature is increased. In conclusion, thethermo gravimetric results show that increasing dilution decreases theoxidation resistance of the coatings.
3.3. X-ray diffraction
Figs. 3 and 4 show the XRD spectra found at the oxidized surface ofthe coatings. The oxide products revealed differences between the twocoatings; however, in each coating the phases produced after oxidationat 800 and 900 °C are similar, although different oxidation weight gainswere measured. According to XRD patterns, the main oxide phase sets
Table 3Nominal chemical composition (wt.%) analyzed at middle thickness of coatings depos-ited using 100 and 128 A arc current.
Elements Ni Fe C Si Al Cr
Coating deposited using 100 A 85.40 8.47 0.30 2.17 0.81 2.86Coating deposited using 128 A 51.09 43.29 1.01 2.54 0.50 1.87
detected at 800 °C and 900 °Cwere: (i) B2O3, SiO2 andOAlB for coatingsproduced with lower dilution and (ii) Fe2O3 (hematite), Fe3O4 (magne-tite) and spinels of NiFe2O4 for higher dilution coatings. The indexationof oxide phases of coating C128 is in agreement with the results fromthe literature, where similar oxide phases were detected after thermalexposure [16,17]. Both coatings also display a high intensity peak closeto 60°, which corresponds to the face centered cubic (Ni-Fe) solid solu-tion (ICDD card 47‐1405). Furthermore, coating C100 displays low in-tensity XRD peaks which correspond to a Ni–Si phase, such as Ni3Si(ICDD card 32‐0699), and a Cr-boride phase Cr5B3 (ICDD card 32‐0278). All these phases are currently identified in this type of micro-structure [9]. The large amount of iron oxide detected at the oxidizedsurface of coating C128 can be attributed to the higher content of ironin this sample. As referred to before, the higher iron content in this coat-ing is attributed to higher dilution of the base material induced by theuse of a higher arc current. Comparing the peaks intensity of coatingC128, oxidized at 800 °C and 900 °C, the amount of oxide productsFe3O4, Fe2O3 and NiFe2O4 increased with temperature. The NiFe2O4 spi-nel phase (ICDD card 10‐0325), identified at the oxidized surface, resultsfrom the reaction of NiO (whichwas also identified as an oxide phase onthe surface of coating C128) with the Fe2O3 phase, as reported byMusićet al. [18]. Furthermore, B-O phases (ICDD card 06‐0297 (B2O3) andICDD card 50‐1505 (B6O)) could be identified from the XRD pattern ofcoating C128.
3.4. Surface morphology
The different chemical compositions of coatings brought about by di-lution led to the formation of different kinds of oxide scales during iso-thermal oxidation, as shown above. The SEM surface morphologies ofthe oxidized coatings at 900 °C are shown in Figs. 5 and 6 for coatingsC100 and C128, respectively. EDS examinations were conducted of theoxidized surface of coatings to characterize their composition. CoatingC100, oxidized at 900 °C, displays two phases: a gray dark phase, identi-fied in Fig. 5a by letter A, and an evenly distributed white phase,indentified by letter B. A magnification of each zone is shown in Fig. 5b
Fig. 2. Isothermal oxidation curves at 800 and 900°C of: a) coating deposited using 100A arc current, b) coating deposited using 128 A arc current.
and c in order to better illustrate the phases. EDS analysis reveals that thegray phase is rich in boron and oxygen, suggesting that it is a boronoxide, as identified by XRD analysis. A similar phase was detected by Liand Qiu [19], and Lavrenko and Gogotsi [20] after oxidation of boroncarbide powder in air, ranging from 500 to 800 °C and, oxidation ofhot-pressed boron carbide in air up to 1500 °C, respectively. They ob-served that even if the isothermal oxidantion temperature exceeds themelting point of the boron trioxide (about 577 °C), after cooling it is pos-sible detect cristaline boron oxide. According to them, at a temperatureclose to 577 °C the boron oxide melts, covering the oxide surface andserving as protective layer. Therefore, oxygen is inhibited from passinginward and metal atoms from passing outward through the oxidelayer. Above a threshold temperature, liquid boron oxide starts to evap-orate, but only after1000 °C does this evaporation becomes significant.Therefore, when the oxidation temperature is below 1000 °C, duringcooling down to room temperature, it is expected that boron oxidewill solidify on top of the oxidized surface of the material. The whitephase appears associated with spalling zones and cracks. The EDS anal-ysis shows that different phases coexist in this zone. A zone rich inboron (see peak ID 2) and another rich in silicon (see peak ID 3) wereidentified, as shown in Fig. 5c. These results matchedwith XRD analysiswhich suggest that the main phases of the oxidized surface are boronand silicon oxides, which are responsible for oxidation resistance ofthe coating due to their protective character [17,19–21]. Moreover, afloret-like structure appears to be frequently associated with thewhite phase, as shown in Fig. 5d. The EDS results (see peak ID 4)show that the elemental composition of this oxide is rich in aluminium.Taking into account the XRD results it can be concluded that the mainphase of the floret-like oxide is OAlB. The morphology and indexationof this phase named aluminum borate whiskers, matches well withthat studied in references [22,23].
However, analysis of the oxidized surface of coating C128 at 900 °Creveals that the oxide morphology is different from that of coatingC100 (see Fig. 6a). It displays amicrostructurewith uniformly distribut-ed irregular ditches and dissimilar phases. Amagnification of themicro-structure is shown in Fig. 6b to make this clearer and more easilycharacterized. The EDS analysis reveals that the phase with irregularditches (peak ID 1) is rich in iron and oxygen, suggesting an ironoxide, as identified in the XRD pattern. Further XRD analysis revealedthat the oxidized surface is composed primarily of Fe2O3 with smallamounts of Fe3O4. This result is in agreement with that of Guo et al.[16]. It is possible to identify other small features such as anotherphase rich in iron (see peak ID 2 relating to the angular particles) forwhich EDS analysis shows the presence of nickel. The combination ofNi with the Fe2O3 phase can give rise to NiFe2O4 spinels which werealso identified by XRD. Furthermore, particles of carbon (dark phase)
could be detected at the oxidized surface of the coating, as shown bypeak ID 3. Finally a phase rich in nickel, oxygen, carbon, silicon andboron (see peak ID 4) was detected at the oxidized surface of coating128 A explaining the Ni–O, and B–O phases identified in the XRDpattern of the specimen.
3.5. Cross‐section of oxidized samples
The microstructures of cross-sections of the oxide scale of coatingsC100 and C128, after oxidation and testing at 900 °C for 2 h, are pres-ented in Figs. 7a and b, respectively. The oxide scale is about 5 μmthick for the specimen coated using the lower arc current and 10 μmthick for that using the higher current. Decarburization of coatingC128 close to the surface is also revealed. Comparing this cross‐sectionwith the oxide scale of the same coating oxidized at 800 °C (Fig. 8) it ispossible to conclude that at this temperature there is no decarburizationof the specimen. Zones of silicon segregationwere found near the inter-face in both coatings (C100 and C128) annealed at either temperature,as shown, as an example, by the EDS analysis performed on the phasemarked with number 1 in Fig. 7c for coating C128 oxidized at 900 °C.
The structure of the oxide scale will be discussed separately for eachcoating, in order tomore clearly illustrate its composition and the role ofthermodynamic stability and transport mechanisms in the establish-ment of layered structures.
Fig. 9 shows SEM images in secondary and backscattered electrons ofcross‐sections of the oxide scales of coatings C100 and C128 at 900 °C.The oxide layers are complex and composed of a mixture of phases.EDS analyses were conducted on the cross‐section of the oxide scales,in order to characterize the different oxides formed during isothermaloxidation. Table 4 summarizes the EDS analyses performed at the pointsmarked in the figures. The values plotted in this table indicate the ratiobetween the peak intensity of each element with the peak intensity ofoxygen for each EDS spectrum. This means that at each point analyzed,a relative increase in this ratio for one particular element is a result ofthe preferential formation of the oxide of that element over the others.Combining SEM imageswith EDS results, plotted in Table 4 it can be con-cluded that the oxide scale of coating C100 ismainly composed of siliconoxide (as the phase identified with point 1 suggests), with smallamounts of a dark phase rich in nickel, silicon, aluminum and boron(see phase identified with point 2) uniformly distributed through thelayer. This agrees well with the XRD pattern and the indexed Si-O,Ni-Si and O-Al-B phases detected. In addition, it is possible to identify aphase rich in boron, normally located at the top of the oxidized surface(point 3), suggesting a boron oxide, as also revealed by XRD. Moreover,a Ni-rich white phase (see point 4) appears at the interface betweenthe coating and the oxide layer, corresponding to the remaining part of
200 F. Fernandes et al. / Surface & Coatings Technology 207 (2012) 196–203
the original source of elements which diffused outwards to form oxidescales.
The oxide scale of coating C128 is different from that of coating C100,being composed of two layers. Combining the XRD analysis, with theSEM-EDS results, the external layer can be identified as Fe2O3, the mainphase containing small features of Fe3O4 and spinel NiFe2O4. On theother hand, the internal layer is composed of a continuous Fe–O richphase (phase identified with point 2) with small evenly distributedamounts of a dark oxide rich in silicon, nickel and iron, as point 3 ofEDS analysis shows. The lower iron content in the internal layer in rela-tion to the external one suggests that the internal Fe–O rich phase re-spects the Fe3O4 oxide. This result is coherent with the oxidation ofFe-based materials and so hematite (Fe2O3) and spinel of NiFe2O4 canbe expected to form in the region of higher oxygen potential while mag-netite (Fe3O4) will form in the lower oxygen potential region [24]. Simi-lar iron oxide layers were observed by Guo et al., after oxidation of Fe–36% Ni bicrystals in air at high temperature [16]. Between these outerand inner layers, a continuous thin layer of a dark phase appears at theinterface, as revealed by SEM. A boron rich phase could be detected atthe top of the outer layer (see point 4), which is in good agreementwith the boron oxide detected by XRD. Finally, clusters of carbon are dis-tributed evenly in the oxide layers. The accumulation of C at the interface
Fig. 5. A) SEM observation of surface morphology of the coating deposited using 100 A arc ctified in A), D) magnification of the zone C identified in C). SEM/EDS spectra of: E) zone 1,
between the coating and the oxide scale (point 5) is a consequence of theoxide layer serving as a diffusion barrier for C coming from the decarburi-zation of the sample mentioned above. Analogous oxide products wereobserved in the cross‐sections of the oxide scale of coating C128 oxidizedat 800 °C.
A good correlation was found between the scale microstructure andthe isothermal oxidation curve of coating C100 oxidized at 900 °C. Theoxidation kinetics of metals and alloys is often controlled by a parabolicrate law at high temperature [25] due to the presence of protective scales.This may well be the case in this specimen, as the evolution of the oxida-tionweight gain is controlled by the silicon oxide, themain phase presentin the oxide scale, as suggested by Fig. 9. At lower temperatures, a boronoxide starts forming on the surface of the specimen due to the high affin-ity of boron to oxygen. However, for higher temperatures, due to the in-tense segregation of silicon through the interface and its great affinity foroxygen, a silicon oxide layer starts to be formed beneath the boron oxide.At the same time small amounts of Ni–Si and O–Al–B phases are formeddue to the diffusion of other elements through the interface. At a temper-ature of approximately 577 °C, boron oxide melts, covering the oxidizedsurface of the coating and inhibiting the transport of species though themelted oxide layer. Above this temperature it is expected that meltedB2O3 starts to evaporate: however, this vaporization does not
urrent, after 2 h oxidation at 900 °C, B) and C) magnification of the zones A and B iden-F) point 2, G) point 3, H) zone 4.
Fig. 7. Back-scattered SEM image of the cross‐section of the oxide scale obtained at 900°C of: A) coating deposited using 100 A arc current, B) coating deposited using 128 A arccurrent. C) Segregation of silicon in the coating deposited using higher arc current, oxidized at 900°C. D) SEM EDS spectra of point 1.
Fig. 6. A) SEM observation of surface morphology of the coating deposited using 128 A arc current, after 2 h oxidation at 900°C, B) magnification of the microstructure in A). SEM EDSspectra of: D) point 1, D) point 2, E) point 3, F) point 4.
Fig. 8. SEM images of the cross‐section of the oxide scale obtained at 800°C of thecoating deposited using 128 A arc current: A) SEM secondary electron (SE) image,B) back-scattered SEM image.
Fig. 9. SEM images of the cross‐section of the oxide scale of the coatings depositedusing100 and1secondary electron (SE) images.
202 F. Fernandes et al. / Surface & Coatings Technology 207 (2012) 196–203
significantly influence the slope of the kinetics curve at 800 and 900 °C,since it only has an important impact at temperatures above 1000 °C[19,20]. Finally, due to the continuous segregation of silicon, its oxidelayer will progressively thicken. A protective layer was thereforeestablished and the reaction rate becomes governed by the rate of speciesdiffusion through this thicker silicon oxide layer, as observed in the cross‐section of the oxide scale of the specimen. As with coating C100, a strongcorrelation can be found between the microstructure of the oxide scaleand the isothermal oxidation curve of coating C128. At 800 °C, parabolicbehavior is displayed; however, at 900 °C, two steps can be detected, seeFig. 2b. The isothermal oxidation curve at first exhibits a linear increase inmass gain and then starts to conform to a parabolic law. The microstruc-tural analysis of the cross-section of the oxide scale revealed that at800 °C there is no decarburization of the coating, which is not whatwas observed at 900 °C. Like the specimen coated using the lower arccurrent, at low temperatures boron oxide starts forming on the surfaceof specimen due to the high affinity of boron to oxygen. Moreover, forhigh temperatures, due to the high iron content, a consistent layer ofFe2O3 with small features of Fe3O4 and NiFe2O4 is formed, which is thick-ened by outward Fe diffusion. At the same time the specimen loses car-bon by decarburization, due to the formation of CO2. However, carbon
28A arc current oxidized at 900°C: A) andC) in back-scattered SEM images, B) andD) in SEM
has increasing difficulties diffusing outwards through the oxide layer andstays either inside or beneath the oxide scale, as shown bymorphologicalanalysis of the cross-section. Therefore, after the first step where theweight gain is partially compensated by C liberation, the oxidationcurve will deviate to a normal parabolic trend. According to theRichardson-Ellingham diagram, Si has higher affinity for O than Ni andFe, and Fe forms Fe3O4 more easily than Ni forms NiO. Therefore, duringoxidation it would be expected that Si-O would be the first layer to beformed. However, its much lower content in the specimen relative toiron, indicates that considerable time is required to form a continuousand protective SiO2 layer, as suggested by Douglass and Armijo [26]. So,after a critical time a layer of SiO2 starts being formed below the Fe2O3
scale, but it never becomes thick enough to determine the oxidation ki-netics. Further, due to the continuous segregation of iron through the sur-face of the specimen and due to the decrease in oxygen content at thecoating's surface as a result of the volume growth of the Fe2O3 layer, acontinuous and consistent Fe3O4 layer starts forming in this zone, withsmall uniformly distributed amounts of the dark phase.
4. Conclusion
This investigation concerned the effect of PTA current variationon the oxidation behavior in air at 800 and 900 °C of coatings of anickel-based hardfacing alloy deposited on gray cast iron using two dif-ferent arc currents (100 and 128 A). Samples were analyzed by thermalgravimetric analysis, x-ray diffraction (XRD) and scanning electron mi-croscopywith energy dispersion spectrometry analysis (SEM-EDS). Thethermo gravimetric results show that increasing the arc current from100 to 128 A decreases the oxidation resistance of coatings, a resultthat can be correlated with the different chemical compositions of thecoatings, promoted by different dilutions of the base material. Thecoating with lower dilution follows a parabolic oxidation weight gainas a function of time, as this behaviour is essentially controlled by thegrowth of a Si–O layer. The coating with higher dilution follows thesame trend at 800 °C. However, at 900 °C two stages in the oxidationcurvewere detected, the firstwith a linear increase inmass gain, a com-promise between outward Fe diffusion through a growing layer ofFe2O3 and the loss of carbon by decarburization and formation of CO2.The second step obeys a parabolic law from the moment that C libera-tion is impeded by a thickened scale consisting of an external Fe2O3
phase, with small amounts of Fe3O4 and NiFe2O4 spinel rich film, and
an internal layer of Fe3O4 film with small evenly distributed amountsof a dark phase rich in silicon, nickel and iron. In summary, an increasein arc current from 100 to 128 A in the deposition of nickel-rich layerson substrates of cast iron by the PTA process is very harmful, as itreduces the oxidation resistance of the coatings at high temperature.
Acknowledgements
The authors wish to express their sincere thanks to the PortugueseFoundation for the Science and Technology (FCT), through COMPETEprogram from QREN and to FEDER, for financial support in the aim ofthe project number “13545”, as well as for the grant SFRH/BD/68740/2010.
References
[1] M. Cingi, F. Arisoy, G. Basman, K. Sesen, Mater. Lett. 55 (2002) 360.[2] L. Luo, S. Liu, J. Li, W. Yucheng, Surf. Coat. Technol. 205 (2011) 3411.[3] R. Vaßen, M.O. Jarligo, T. Steinke, D.E. Mack, D. Stöver, Surf. Coat. Technol. 205
(2010) 938.[4] C. Guo, J. Zhou, J. Chen, J. Zhao, Y. Yu, H. Zhou, Wear 270 (2011) 492.[5] K. Gurumoorthy, M. Kamaraj, K.P. Rao, A.S. Rao, S. Venugopal, Mater. Sci. Eng., A
456 (2007) 11.[6] W. Li, Y. Li, C. Sun, Z. Hu, T. Liang, W. Lai, J. Alloys Compd. 506 (2010) 77.[7] A. Gatto, E. Bassoli, M. Fornari, Surf. Coat. Technol. 187 (2004) 265.[8] K. Siva, N. Murugan, R. Logesh, Int. J. Adv. Manuf. Technol. 41 (2009) 24.[9] F. Fernandes, B. Lopes, A. Cavaleiro, A. Ramalho, A. Loureiro, Surf. Coat. Technol.
205 (2011) 4094.[10] V. Balasubramanian, A.K. Lakshminarayanan, R. Varahamoorthy, S. Babu, J. Iron
Steel Res. Int. 16 (2009) 44.[11] C.T. Liu, J. Ma, X.F. Sun, J. Alloys Compd. 491 (2010) 522.[12] W.Z. Li, Y.Q. Li, Q.M. Wang, C. Sun, X. Jiang, Corros. Sci. 52 (2010) 1753.[13] K.-c. Zhou, L. Ma, Z.-y. Li, Trans. Nonferrous Met. Soc. China 21 (2011) 1052.[14] G.R. Wallwlork, Rep. Prog. Phys. 39 (1976) 401.[15] E.O. Ezugwu, Z.M. Wang, A.R. Machado, J. Mater. Process. Technol. 86 (1998) 1.[16] X. Guo, K. Kusabiraki, S. Saji, Oxid. Met. 58 (2002) 589.[17] A.R. Lashin, O. Schneeweiss, M. Svoboda, Oxid. Met. 69 (2008) 359.[18] S. Musić, S. Popović, S. Dalipi, J. Mater. Sci. 28 (1993) 1793.[19] Y.Q. Li, T. Qiu, Mater. Sci. Eng., A 444 (2007) 184.[20] V.A. Lavrenko, Y.G. Gogotsi, Oxid. Met. 29 (1988) 193.[21] Y. Wu, Y. Niu, Scr. Mater. 53 (2005) 1247.[22] H.K. Lee, S. Zerbetto, P. Colombo, C.G. Pantano, Ceram. Int. 36 (2010) 1589.[23] I.E. Gönenli, T.A. C., Powder Diffract. 15 (2000) 104.[24] V.B. Trindade, B. Rodrigo, B.Z. Hanjari, S. Yang, U. Krupp, C. Hans-Jürgen, Mater.
Res. 8 (2005) 365.[25] Y. Nakamura, Metall. Mater. Trans. A 6 (1975) 2217.[26] D.L. Douglass, J.S. Armijo, Oxid. Met. 2 (1970) 207.
Annex C
97 Filipe Daniel Fernandes
Annex C
F. Fernandes, A. Ramalho, A. Loureiro, A. Cavaleiro, Wear resistance of a nickel-
based coating deposited by PTA on grey cast iron, International Journal of Surface
Science and Engineering, 6 (2012) 201-213.
Int. J. Surface Science and Engineering, Vol. 6, No. 3, 2012 201
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron
Filipe Fernandes* CEMUC – Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, Portugal and Intermolde, Rua de Leiria, 95, Apartado 103, 2431-902 Marinha Grande, Portugal Fax: +(351) 239-790-701 E-mail: [email protected] *Corresponding author
Amilcar Ramalho, Altino Loureiro and Albano Cavaleiro CEMUC – Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, Portugal Fax: +(351)-239-790-701 E-mail: [email protected] E-mail: [email protected] E-mail: [email protected]
Abstract: The moulds for the production of glass bottles, made of cast iron, are subjected to very severe conditions of wear during use. Thus, it is essential to understand the wear mechanisms involved, in order to increase the equipment life. The present study intends to study the abrasion resistance of moulds submitted for long time at working conditions. The investigation was conducted on nickel-based coatings deposited by plasma transferred arc (PTA) on grey cast iron, using different arc currents. Micro-scale ball cratering abrasive wear test was used to evaluate the tribological properties of the as-deposited and heat treated coatings. The results show that micro-scale abrasion tests induce grooving and pitting wear mechanisms. Increasing the arc current decreased the hardness of the coatings and, consequently, their wear resistance. The hardness and wear resistance of the coatings was improved by increasing heat treatment holding time.
Reference to this paper should be made as follows: Fernandes, F., Ramalho, A., Loureiro, A. and Cavaleiro, A. (2012) ‘Wear resistance of a nickel-based coating deposited by PTA on grey cast iron’, Int. J. Surface Science and Engineering, Vol. 6, No. 3, pp.201–213.
202 F. Fernandes et al.
Biographical notes: Filipe Fernandes is a PhD student in the Mechanical Engineering Department of the University of Coimbra. He obtained his MSc in 2009. He is currently dedicated to the surface engineering field applied to the moulds for the glass industry. He is focused on the study of new methods to protect the surface of those components. His research interests include thermal spraying of coatings (PTA, HVOF, APS and PVD), oxidation and wear resistance of materials suitable for high temperature applications.
Amilcar Ramalho received his PhD in Mechanical Engineering from Universidade de Coimbra, Portugal in 1994. He is currently a Professor of Maintenance and Mechanical Vibrations. His research interests include friction and wear of materials and components, fretting, tribotesting and biotribology. He is the co-author of more than 150 papers in international journals and conferences. He belongs to the editorial board and collaborates as a reviewer of several international journals.
Altino Loureiro is currently Associate Professor of Welding and Related Technologies in the University of Coimbra (UC) and Researcher from the Center of Mechanical Engineering at the same university (CEMUC). His research interests are focused on technological and metallurgical aspects of welding and thermal spaying processes. He is the author or co-authored over 150 articles in international journals and conferences and reviewer of several international scientific and technical journals. His is also coordinated and collaborated on more than a dozen of research projects and is a consultant to companies in the field of metal construction.
Albano Cavaleiro is full Professor in the Mechanical Engineering Department of the University of Coimbra. He is dedicated to the surface engineering field for more than 25 years. He is mainly concerned with the deposition and characterisation of sputtered coatings for mechanical applications, such as, low friction and self-lubricant, high temperature resistant, oxidation and corrosion protective and DLC. He published more than 150 papers in international peer reviewed journals and presented more than 30 invited talks in international scientific events.
1 Introduction
Cast iron is commonly used in the production of moulds and accessories for glass industry. Wear and oxidation are the main failure mechanisms that limit the surface life of moulds (Cingi et al., 2002). Therefore, it is necessary to protect mould surfaces with hard and wear resistant materials.
The plasma transferred arc (PTA) process has been usually applied to coat the surfaces of moulds and, especially, their edges where the wear is higher. It produces very high quality deposits, offering optimal protection with minimal thermal distortion of the parts, and provides high deposition rate (Gatto et al., 2004). It allows very precise deposit layers of complex alloys on mechanical parts that are subjected to harsh environments, significantly extending their service life (Siva et al., 2009). However, the hardness and the wear resistance of the coatings are strongly influenced by the dilution of the substrate promoted by the PTA process. Low dilution provides coatings with similar chemical composition to the metal powder added, condition to have an improved wear and
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron 203
corrosion resistance (Balasubramanian et al., 2009). Therefore, it is essential to control the relevant process parameters to minimise the dilution.
A significant number of coating materials, such as cobalt, iron and nickel alloys, can be applied by PTA process to extend the working life of tribological components subjected to wear. In glass industry nickel alloys gain extremely important role in protection of surfaces, due to their excellent performance under conditions of abrasion and corrosion at elevated temperature. The microstructure and wear behaviour of nickel-based hard-facing alloy deposits have been studied using various alloy compositions and different substrates. Nickel-based alloys are usually strengthened by alloying elements which improve their wear properties (Ezugwu et al., 1998; Chang et al., 2010). Technical literature reports a lot of tribological studies done by pin-on-disc tests in deposits of Ni alloys on stainless and carbon steels, though the sliding contact conditions of pin-on-disc do not reproduce the interaction conditions between melted glass and the surface of the mould. Because of that, the ball cratering wear test was selected as the first step to investigate the micro-abrasion resistance of the coatings at room temperature.
Guo et al. (2011) studied the high temperature wear resistance and wear mechanism of NiCrBSi and NiCrBSi/WC-Ni coatings deposited on stainless steel by laser cladding. They found that the wear resistance of the coatings is much greater than that of stainless steel when sliding against Si3N4 counterpart ball. Furthermore, NiCrBSi/WC-Ni composite coating showed better wear resistance at high temperature than NiCrBSi coating. Kesavan and Kamaj (2010), who worked with a Colmomoy 5 hard-facing alloy deposited by PTA on a stainless steel, studied the room and high temperature wear behaviour using pin-on-disc wear tests. They showed that the wear resistance of the coatings improved significantly with increasing test temperature. Further, they observed different operating wear mechanisms depending on the sliding and testing temperature. In turn, Gurumoorthy et al. (2007) studied the sliding wear of a nickel hard-facing alloy AWS NiCr-B deposited on the same stainless steel at three different temperatures (room temperature, 300°C and 500°C), and the influence of an ageing treatment on its microstructure, hardness and wear behaviour. The investigation revealed a significant weight loss at room temperature and an abrupt decrease in mass loss at high temperature. However, after ageing treatment they demonstrated that there is no significant change in the weight loss or the wear behaviour at room and high temperature.
Since wear and oxidation are the main drawbacks that limit the surface life of glass moulds, the aim of this research is to study the effect of PTA current on the microstructure, hardness and three-body abrasion behaviour of a nickel-based hard-facing alloy coating deposited on grey cast iron. Ageing studies were also conducted to evaluate the influence of temperature holding time on the microstructure, hardness and tribological properties of the coating processed with the highest current.
2 Experimental procedure
A nickel-based hard-facing alloy named Colmonoy 215 (from Colmonoy Company) was deposited on flat surfaces of grey cast iron blocks by PTA process. Before coating, each cast iron block was preheated at 480°C, in order to reduce the thermal shock and cooling rate to avoid cracks in both the coating and heat-affected zone. The same process
204 F. Fernandes et al.
parameters were used for all coatings except arc current, which was 100 A for specimen 1, 128 A for specimen 2 and 140 A for specimen 3. The composition of cast iron and metal powder is given in Table 1 and the process parameters used in Table 2. After coating, the blocks were left to cool to room temperature. Samples were removed from each block, transversely to welding direction, for further analysis. Ageing studies were carried out at 500°C for 5, 10 and 20 days on specimen 3. Table 1 Nominal chemical composition (wt.%) of substrate and hard-facing alloy
Base material C Mn Si P S Cr Ni Mo V Ti Fe
Grey cast iron 3.60 0.60 2.00 < 0.20 < 0.04 < 0.20 < 0.50 0.50 0.10 0.20 Balance Hard-facing C Cr Si B Fe Al F Co Ni Ni-alloy 0.14 2.45 2.56 0.86 1.08 1.30 0.01 0.08 Balance
Table 2 Deposition parameters for PTA weld surfacing.
Microstructural analysis was done in cross section of the samples. Optical and scanning electron microscopies (OM and SEM) were used to characterise the microstructure of the coatings. Vickers micro-hardness profiles were measured in the cross section of treated samples using a load of 5 N.
Micro abrasive wear tests were carried out in as-deposited and heat treated surfaces at room temperature in a ball cratering devices. Figure 1 shows a schematic diagram of the equipment used. In this test a specific normal load is applied to press a rotating ball into the flat surface of the coated sample in the presence of a slurry suspension of abrasive. The abrasive slurry, agitated by a magnetic stirrer, was continuous and gravitationally drip fed onto the rotating ball to produce a spherical cup depression. Preliminary tests using different contact conditions were done to reproduce the interaction conditions in the surface of the moulds which have been in service for a long time. The wear mechanism was studied by SEM. High purity silica SS40 with angular particles of averaging of 3.10 μm in size was used as an abrasive to replicate the effect of the melted glass. The slurry used was prepared with a concentration of 103 g per 1 × 10–4 m3 of distilled water. A steel ball bearing (AISI 52100) with 0.0254 m was used. Several scars were made in each sample with different durations, 100, 200, 300, 400 and 500 turns, respectively 7.98, 15.96, 23.94, 31.92, 39.90 meters of sliding distance for all the specimens, under the action of a normal load of 0.1 N. In order to minimise the scatter of results, previously to starting the tests, the ball surface was etched with chloridric acid (33%) during ten
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron 205
minutes. The dimensions of spherical depressions were measured in order to calculate the wear volume, which is given by the equation (1). In this equation, R represents the ball radius and b the crater chordal diameter of the spherical cup depression. The b value was evaluated using the average of two perpendicular measurements performed at the wear scar chordal diameter. Those values were measured with a Mitutoyo Toolmaker Microscope with x-y micrometer table.
( )4V π b (64 R)= × × × (1)
In order to obtain the specific wear rate, to quantify the wear behaviour, Archard’s law was applied, see equation (2). This is the most applied law to determine a parameter to quantify the wear behaviour of materials. In this equation, P is the normal load, l is the sliding distance and k is the specific wear rate.
V k P l= × × (2)
To quantify the specific wear rate a linear approach of Archard’s model was applied and a statistical analysis was used to estimate the error of measurements, as described elsewhere (Ramalho, 2010).
Figure 1 Schematic diagram of the micro-scale abrasion tests
3 Results
3.1 Microstructure
Figure 2 shows the microstructure of the coatings performed using different arc currents [Figures 2(a), 2(b) and 2(c)], as well as the fusion boundary of specimen 3 [Figure 2(d)], and SEM pictures of specimen 3 before and after heat treatment [Figures 2 (e) and 2(f)]. The typical microstructure of deposits consists of dendrites of Ni-Fe solid solution phase with columnar morphology oriented along the direction of heat flow, with C-flakes (dark floret-like structures) randomly distributed. However, near the interface it was observed the presence of a cellular microstructure finer than the dendrite structure with diminution of the size of C-flakes, as illustrated in Figure 2(d) for specimen 3. This behaviour could be explained by the higher solidification rates involved in the boundary because of the
206 F. Fernandes et al.
efficient thermal exchange ensured by the high volume ratio substrate/coating of the base material, as suggested by Gatto et al. (2004) and Navas et al. (2006). Moreover, the presence of the ‘dark’ phase on the coatings cannot be explained by the chemical composition of the metal powder added, so it can be only attributed to the high dilution of cast iron induced by the PTA process. The dilution was evaluated by the ratio between the area of the melted base material and the total area of the melted zone, both measured in the cross section of the coatings. Dilutions of 28%, 50% and 59% were measured respectively for specimen 1, 2 and 3. These results agree with the observed increasing proportion of C-flakes for higher arc currents, as can be seen in Figures 2(a), 2(b) and 2(c). Therefore, the chemical composition of the coatings is largely influenced by the dilution of cast iron. A detailed characterisation of this type of coatings can be found in a previous publication of the authors (Fernandes et al., 2011).
Figure 2 Microstructure of (a) specimen 1, (b) specimen 2, (c) specimen 3, (d) interface of specimen 3 and SEM pictures of sample 3 (e) in non-heat-treated condition and (f) in heat treated condition during ten days
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron 207
Table 3 SEM chemical analysis of the deposit and base material of specimen 3
Figure 3 SEM magnification of microstructure of the coating 3 (a) in non-heat treated condition and (b) in heat-treated condition after ten days exposed at 500°C and SEM EDAX spectra of (c) point 1, (d) point 2, and (e) point 3 (see online version for colours)
(c) (d)
(e)
Table 3 shows the chemical composition in area measured by SEM-EDS at various zones on the coating and the substrate of specimen 3. Zone 1 and 2 are located close to the coating surface, zone 3 at the mid-thickness of the coating, zone 4 in the coating close to the fusion boundary and zone 5 in the bulk cast iron. This analysis showed that PTA process promotes an evenly distribution of the elements inside the coating. The same observation was done in specimens 1 and 2. Due to the dilution, the amount of iron in the
208 F. Fernandes et al.
coating increases reducing its nickel content. Ezugwu et al. (1998) reported that the increase of iron content in the coatings tends to decrease their oxidation resistance, because of the formation of a less adherent oxide scale.
Figures 2(e) and 2(f) reveals that the heat treatment induces significant changes in the original microstructure of specimen 3. It promotes the formation of a white phase on the grain boundaries. Magnifications of the two microstructures are shown in Figure 3. SEM-EDS spectra analyses were conducted on the coatings to give a detailed characterisation of their microstructure. From the EDS analysis, it can be stated that the grain boundaries of the non-heat-treated sample is formed by two phases: a light grey and a dark grey rich in chromium and silicon, respectively. Inversely, the specimen in heat-treated condition is formed by the light grey phase as observed in the non-heat-treated sample and a new phase (white phase) which is replacing the dark phase. The amount of the white phase grows with increasing time of heat treatment. According to Fernandes et al. (2011), who studied a nickel alloy deposited on grey cast iron, the dark grey phase is a Ni3Si phase. The EDS spectrum reveals that both phases (the dark and the white) are rich in Ni-Si. However, the white phase presents a higher content of phosphorous on the heat treated sample suggesting a dispersion of Ni-P precipitates on the grain boundaries, as observed by Sahoo and Das (2011) in electroless nickel coatings.
Figure 4 Hardness profile across the interface of (a) as-deposited specimens and (b) specimen 3 before and after heat treatment (see online version for colours)
(a) (b)
3.2 Hardness
The micro-hardness profiles of the different as-deposited and heat treated specimens are shown respectively in Figures 4(a) and 4(b). Figure 4(a) reveals that the hardness through the coatings is approximately constant and tends to decrease with increasing arc current. The dilution induced by the PTA process can explain this behaviour. The partial melted zone (PMZ) displays the highest hardness. According to Pouranvari (2010), who studied the weldability of grey cast iron using nickel-based filler metals, the highest hardness observed in PMZ is due to the formation of hard and brittle phases during surfacing, such as martensite and white cast iron. The heat treatment promotes the progressive hardening of the coatings with increasing thermal exposure time [see Figure 4(b)]. Some scatter of the hardness values can be observed too. Harsha et al. (2008), who studied the influence of CrC in a nickel flame sprayed coatings, observed similar behaviour. According to them, the scatter in the values is due to the presence of a harder eutectic network around the cells. As described before, heat treatment promotes the formation of Ni-P precipitates
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron 209
on the grain boundaries. The greater hardness observed on the ageing treated samples can be attributed to the presence of these precipitates similarly to what is currently observed in Ni-P electroless coating when annealed at increasing temperatures (Sahoo and Das, 2011). This type of coatings is a supersaturated alloy in the as-deposited state and can be strengthened by the precipitation of nickel phosphide crystallites. The phosphides act as barriers for dislocation movement and improve the mechanical strength of the coatings.
3.3 Wear behaviour of coatings
Micro-scale abrasion tests were performed on the coatings at room temperature to predict the effect of the different process parameters on their wear behaviour. The goal in a first step was to study the abrasion resistance present in the surface of in-service moulds, in order to quantify the wear loss of coatings. The response of materials under micro abrasion tests depends on the nature of the motion of particles in the contact zone, which determines the stress state arising near the contact. Three-body wear mechanism (rolling or pitting) and two-body wear mechanism (grooving), or a combination of both mechanisms could be detected. Rolling occurs when the abrasive particles are able to free rotate between the sphere and the material in test; on the other hand, grooving occurs if the particles follow the rotation of the sphere, scratching the specimen surface (Ramalho et al., 2011; Gant and Gee, 2011). In spite of this general tendency to the occurrence of three-body or two-body abrasion, the characteristics of the materials will influence the process; mild materials will stabilise the grooving mode.
Figure 5 Evolution of the wear volume as function of the parameter ‘normal load × sliding distance’ of (a) as-deposited coatings and (b) specimen 3 before and after thermo exposure at 500°C during several days (see online version for colours)
(a) (b)
The results of the micro-scale abrasion tests from specimens in as-deposited and heat treaded conditions are plotted respectively in Figures 5(a) and 5(b), as the material volume loss as a function of the product of the sliding distance (l) by the normal load (P). The reliability analysis of the data is presented in Table 4 and the confidence interval for the specific wear rate plotted in Figure 6. For each sample, a straight linear trend was fitted to the experimental points in order to obtain the specific wear rate, which corresponds to its slope [see equation (2)]. From Figure 5(a), it is possible to conclude that specimen 1 exhibits slightly lower specific wear rate, therefore it is the more resistant to wear. This result is in accordance with its greater hardness when compared to the others as-deposited coatings. Specimens 2 and 3 displays similar wear volumes, which is
210 F. Fernandes et al.
again attributed to their similar hardness profiles. Figure 5(b) reveals that specimens exposed during five and ten days at 500°C displayed similar specific wear rate to the non-heat-treated specimen. However, the specimen exposed during 20 days exhibits lower specific wear rate, therefore, it is the most resistant to wear. This last result is again in accordance with the higher hardness of this sample after the heat treatment. Table 4 Results of the linearisation analysis of samples in as-deposited and heat treated
conditions
Specific wear rate, k(mm3/N m)
Average STD Confidence interval 90% r2
Specimen 1 2.83 × 10–3 2.65 × 10–4 2.20 × 10–3 to 3.45 × 10–3 0.974 Specimen 2 3.44 × 10–3 3.78 × 10–4 2.55 × 10–3 to 4.33 × 10–3 0.965 Specimen 3 3.45 × 10–3 2.14 × 10–4 2.94 × 10–3 to 3.95 × 10–3 0.989 After 5 days at 500°C 3.69 × 10–3 5.57 × 10–5 3.56 × 10–3 to 3.83 × 10–3 0.999 After 10 days at 500°C 3.73 × 10–3 1.91 × 10–4 3.28 × 10–3 to 4.18 × 10–3 0.992 After 20 days at 500°C 2.58 × 10–3 3.18 × 10–4 1.83 × 10–3 to 3.33 × 10–3 0.956
Figure 6 Confidence interval for a confidence level of 90%, for the specific wear rate of specimens in as-deposited and heat treated conditions (see online version for colours)
In order to identify the wear mechanisms and to compare the worn surfaces with those observed in the surface of moulds, SEM examinations were carried out. Figure 7(a) displays a SEM micrograph of the surface of a coating of a mould after long time in service. From this picture is possible to see that the surface failure regions display pits and cracks. Considering the in-service conditions the surface damage probably was induced by synergetic effects of several mechanisms, namely thermal fatigue, corrosion and abrasion. However, pitting is the main active failure mechanism of the moulds. Figures 7(b) and 7(c) illustrates spherical cup depressions induced by the micro-scale abrasion tests on the specimen 3, before and after holding time for 20 days, produced after 500 ball rotations. Both scars display similar wear behaviour with two modes of wear: three-body abrasion (pitting) and two-body abrasion (grooving). The two-body abrasion mode is preferentially localised in the area of higher pressure (input slurry area)
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron 211
and three-body abrasion mode is localised in the area of less pressure. Figure 8 shows these two modes of wear at higher magnification. The same wear mechanisms were observed in the other specimens in as-deposit and heat treated conditions. Although the scars displayed two modes of wear it can be stated that the region corresponding to three-body mode reproduces closely the wear mechanism present in the moulds surface. However, one should keep in mind that these results do not reproduce with accuracy the in-service condition of the moulds, since in service they are exposed to other damage mechanisms such as thermal and mechanical fatigue, oxidation and corrosion at high temperature.
Figure 7 SEM micrograph (a) of the surface of a coating of a mould after long exposure to high temperature, (b) scar with 500 rotations induced by the micro-scale abrasion tests on specimen 3, and (c) scar with 500 rotations induced by the micro-scale abrasion tests on specimen 3 after thermal exposure for 20 days (see online version for colours)
Figure 8 SEM micrograph of (a) three-body abrasion and (b) two-body abrasion
212 F. Fernandes et al.
4 Conclusions
Micro-scale abrasion testing was used to study the interaction conditions between melted glass and the surface of the moulds. The influence of the arc current used in PTA process, as well as the effect of heat treatments, on the microstructure, hardness and wear resistance of the coatings was analysed. The increase in arc current increased the dilution of the base material changing the composition and microstructure of the deposits and reducing their hardness. Furthermore, the wear loss of material increased with increasing arc current. The micro-scale abrasion tests displayed two modes of wear: three-body abrasion (pitting) and two-body abrasion (grooving). Although the scars displayed two modes of wear it can be stated that the wear mechanism of the abrasion test reproduces with success the wear mechanism present in the moulds surface. The heat treatment performed on the sample 3 promotes increasing hardness of the coating with thermal exposure time. The enhanced hardness of the coatings is partially attributed to Ni-P precipitates on the grain boundaries. Specimen 3 heat treated for 20 days showed the best resistance to wear.
Future experiments will be devoted to study the high temperature abrasion resistance of the coatings, in order to reproduce the in service conditions of moulds.
Acknowledgements
The authors would like to thank the company ‘Intermolde’, for supplying the coated samples, and the Portuguese Foundation for the Science and Technology (FCT), through COMPETE programme from QREN and to FEDER, for financial support in the aim of the project number ‘013545’, as well as for the grant SFRH/BD/68740/2010.
References Balasubramanian, V., Lakshminarayanan, A.K., Varahamoorthy, R. and Babu, S. (2009)
‘Application of response surface methodolody to prediction of dilution in plasma transferred arc hardfacing of stainless steel on carbon steel’, International Journal of Iron and Steel Research, Vol. 16, No. 1, pp.44–53.
Chang, J.H., Chang, C.P., Chou, J.M., Hsieh, R.I. and Lee, J.L. (2010) ‘Microstructure and bonding behavior on the interface of an induction-melted Ni-based alloy coating and AISI 4140 steel substrate’, Surface and Coatings Technology, Vol. 204, No. 2, pp.3173–3181.
Cingi, M., Arisoy, F., Basman, G. and Sesen, K. (2002) ‘The effects of metallurgical structures of different alloyed glass mold cast irons on the mold performance’, Materials Letters, Vol. 55, No. 6, pp.360–363.
Ezugwu, E.O., Wang, Z.M. and Machado, A.R. (1998) ‘The machinability of nickel-based alloys: a review’, Journal of Materials Processing Technology, Vol. 86, Nos. 1–3, pp.1–16.
Fernandes, F., Lopes, B., Cavaleiro, A., Ramalho, A. and Loureiro, A. (2011) ‘Effect of arc current on microstructure and wear characteristics of a Ni-based coating deposited by PTA on gray cast iron’, Surface and Coatings Technology, Vol. 205, No. 16, pp.4094–4106.
Gant, A.J. and Gee, M.G. (2011) ‘A review of micro-scale abrasion testing’, Journal of Physics D: Applied Physics, Vol. 44, No. 7, 073001.
Gatto, A., Bassoli, E. and Fornari, M. (2004) ‘Plasma transferred arc deposition of powdered high performances alloys: process parameters optimisation as a function of alloy and geometrical configuration’, Surface and Coatings Technology, Vol. 187, Nos. 2–3, pp.265–271.
Wear resistance of a nickel-based coating deposited by PTA on grey cast iron 213
Guo, C., Zhou, J., Chen, J., Zhao, J., Yu, Y. and Zhou, H. (2011) ‘High temperature wear resistance of laser cladding NiCrBSi and NiCrBSi/WC-Ni composite coatings’, Wear, Vol. 270, Nos. 7–8, pp.492–498.
Gurumoorthy, K., Kamaraj, M., Rao, K.P., Rao, A.S. and Venugopal, S. (2007) ‘Microstructural aspects of plasma transferred arc surfaced Ni-based hardfacing alloy’, Materials Science and Engineering: A, Vol. 456, Nos. 1–2, pp.11–19.
Harsha, S., Dwivedi, D. and Agarwal, A. (2008) ‘Influence of CrC addition in Ni-Cr-Si-B flame sprayed coatings on microstructure, microhardness and wear behaviour’, The International Journal of Advanced Manufacturing Technology, Vol. 38, No. 1, pp.93–101.
Kesavan, D. and Kamaraj, M. (2010) ‘The microstructure and high temperature wear performance of a nickel base hardfaced coating’, Surface and Coatings Technology, Vol. 204, No. 24, pp.4034–4043.
Navas, C., Colaço, R., Damborenea, J. and Vilar, R. (2006) ‘Abrasive wear behaviour of laser clad and flame sprayed-melted NiCrBSi coatings’, Surface and Coatings Technology, Vol. 200, No. 24, pp.6854–6862.
Pouranvari, M. (2010) ‘On the weldability of grey cast iron using nickel based filler metal’, Materials & Design, Vol. 31, No. 7, pp.3253–3258.
Ramalho, A. (2010) ‘A reliability model for friction and wear experimental data’, Wear, Vol. 269, Nos. 3–4, pp.213–223.
Ramalho, A., Rodríguez, J. and Rico, A. (2011) ‘Microabrasion resistance of nanostructured plasma-sprayed coatings’, Int. J. Surface Science and Engineering, Vol. 5, No. 4, pp.250–260.
Sahoo, P. and Das, S.K. (2011) ‘Tribology of electroless nickel coatings – a review’, Materials & Design, Vol. 32, No. 4, pp.1760–1775.
Siva, K., Murugan, N. and Logesh, R. (2009) ‘Optimization of weld bead geometry in plasma transferred arc hardfaced austenitic stainless steel plates using genetic algorithm’, The International Journal of Advanced Manufacturing Technology, Vol. 41, No. 1, pp.24–30.
Annex D
113 Filipe Daniel Fernandes
Annex D
F. Fernandes, A. Ramalho, A. Loureiro, A. Cavaleiro, Mapping the micro-
abrasion resistance of a Ni-based coating deposited by PTA on gray cast iron, Wear,
292-293 (2012) 151-158.
Wear 292–293 (2012) 151–158
Contents lists available at SciVerse ScienceDirect
Wear
0043-16
http://d
n Corr
E-m
journal homepage: www.elsevier.com/locate/wear
Mapping the micro-abrasion resistance of a Ni-based coating depositedby PTA on gray cast iron
F. Fernandes n, A. Ramalho, A. Loureiro, A. Cavaleiro
CEMUC, Department of Mechanical Engineering, University of Coimbra, Rua Luıs Reis Santos, 3030-788 Coimbra, Portugal
Micro-scale abrasion was used to characterize abrasion resistance of a nickel-based hardfacing alloy
deposited by Plasma Transferred Arc on gray cast iron using silica as abrasive agent. In order to
investigate the occurrence of different abrasion mechanisms, several test conditions were used,
namely: different silica abrasive contents and a range of normal loads and test durations. A scanning
electron microscope (SEM) was used to study the morphologies of the spherical cup-shaped depres-
sions induced by different test conditions. The results are discussed in terms of the effect of the
dominant wear mechanisms on the abrasion resistance and the influence of the test conditions on the
mechanism transition. The wear results lead to the conclusion that the specific wear rate essentially
depends on the wear mechanisms (rolling or grooving) involved and not on the tests conditions
employed, since these do not produce changes in the wear mechanism.
& 2012 Elsevier B.V. All rights reserved.
1. Introduction
The micro-scale abrasion wear test is a suitable experimentalmethod for evaluating and comparing the abrasive wear perfor-mance of a wide variety of materials. This technique has beensuccessfully applied to characterize metallic and non-metallic,bulk and coating materials, under the influence of a wide diversityof abrasive slurries [1,2]. It is well known that in the industry, themain wear problems in components are related to abrasion,which could co-exist, whether associated or not, with other formsof wear, such as adhesion or tribo-corrosion [3–5]. The usefulnessof the laboratory characterization depends on the ability of thetest conditions to reproduce real working conditions as closely aspossible.
In the micro-scale abrasion test, three distinct wear modeshave been identified, three-body or rolling abrasion, two-body orgrooving abrasion and the ridging wear mechanism [1,2,5–8].Additionally transitions between these wear modes could beobserved (rolling to grooving, grooving to ridging and ridging torolling); these are referred to as ‘‘mixed-mode’’ wear mechan-isms. Rolling abrasion is characterized by a hertzian type contactwhich occurs when the abrasive particles are able to freely rotatebetween the sphere and the material in test, producing multipleindentations, while grooving occurs if the particles remain fixedto the sphere and scratch the specimen surface. On the other
ll rights reserved.
x: þ351 239 790 701.
nandes).
hand, ridging wear mechanism is produced when the abrasiveparticles disappear in the contact region, the contact being madedirectly between the sphere and the specimen tested. Therefore,the ridging wear mode should be considered as transition fromabrasion to metal–metal sliding. Furthermore, according to someauthors [8,9], associated with the complex regimes above men-tioned, tribo-chemical interactions must be taken into account inthe wear calculations specially at high applied loads. This phe-nomenon is known to enhance wear resistance with an increasein applied load due to oxide formation on the surface.
In order to use micro-abrasion tests to evaluate the abrasivewear resistance of a system, either three-body or two-bodyabrasions need to be considered in the study; each wear modeis characterized by significantly different values of specific wearrate, which can differ by more than one order of magnitude [1,10].Some authors stated that, in micro-scale abrasion tests, the rollingabrasion wear mode leads to results that are more easily repro-duced [11,12]. Williams and Hyncica [13,14] discussed the pos-sibility of applying the continuum mechanics abrasion model toexplain the wear transitions. They verified that the quantitativemodels can be only applied to processes or materials where theboundary conditions are well established, which is not the casefor the micro-scale abrasion test. Nevertheless, Williams andHyncica suggested that the transition from rolling to groovingmust be a function of the critical thickness of the abrasive slurryin the contact. Following this suggestion, Adachi and Hutchings[6] used an indentation model to analyze the grooving wearconditions and they proposed a dimensionless parameter forthe severity of contact (S0), to be correlated with the transition
Fig. 1. Typical microstructure of the as-deposited coating.
Fig. 2. Schematic diagram of the of the micro-abrasive wear equipment used.
F. Fernandes et al. / Wear 292–293 (2012) 151–158152
conditions. The ratio of hardnesses between the specimen and theball, Hs/Hb, was identified as the key parameter in the transitionfrom three-body to two-body abrasion. The results showed thatwith increasing severity of contact (S0) a transition from uniformthree-body to two-body abrasion occurs through a mixed threeand two body abrasion regime.
The technical literature reports several studies undertakenusing micro-scale abrasion equipment and describes how thedissimilar test conditions (volume fraction of abrasive in theslurry, abrasive material, applied load, ball and specimen materi-als, sliding distance and surface condition of the ball) influencethe dominant wear modes of different materials [1,6,10,15–18].The main objective of these studies is to track the dominant wearmodes created with different conditions to produce an associatewear-mode map, to serve as data to predict the conditions toensure either three-body or two-body abrasion. With regard tothe type of abrasive material to be used in micro-scale abrasiontests, the BS-EN 1071-6: 2007 standard recommends the use ofSiC F1200 abrasive; however, dissimilar abrasive materials withdifferent grain sizes have been used and, normally, the studiesalso report the use of diamond and Al2O3. It has been shown thatthe relative wear rates of the materials depended strongly uponthe abrasive type selected (abrasive shape and hardness) [19].Nevertheless, there are few references regarding the use of silicaas abrasive material. Because silica is the most common abrasivematerial, therefore responsible for the majority of practical abra-sion problems, it is important to study the test conditions whichproduce different wear mechanisms using silica abrasive slurry.
Thus, the purpose of this investigation was to study the testconditions which produce rolling and grooving wear modes withsilica slurry. An associate wear mode-map should be produced touse as data to ensure either one or the other abrasion wearmechanism’s intended further aim was to show that the specificwear rate essentially depends on the wear mechanisms (rolling orgrooving) involved and not on the test conditions employed, sincethese do not produce changes in the wear mechanism. The weartests were conducted in fixed ball equipment on a nickel-basedhardfacing alloy coating deposited by PTA – Plasma TransferredArc on gray cast iron. Regimes of wear behavior were identified byscanning electron microscope (SEM) and the results were ana-lyzed with a focus on the effects of wear mechanisms.
2. Experimental procedure
In this study, micro-scale abrasion equipment was used tostudy the influence of the test conditions on the dominant wearmodes of a nickel-based hardfacing coating. The coating analyzedwas produced by plasma transferred arc process (PTA) on a flatsurface of gray cast iron using a 128 A arc current. Fig. 1 showsthe typical microstructure of the as-deposited coating. As can beobserved, the coating has a relatively dense microstructure, freeof microcracks and few solidification voids with dendrites ofNi–Fe solid solution phase aligned along the direction of heatflow. Furthermore, it displays C-flakes (dark-floret like structures)randomly distributed in the matrix. The chemical composition ofthe materials, the conditions and the procedure adopted for thedeposition and the characterization of the coatings as well as theirmain physical structural and mechanical properties can be foundin a previous publication of the authors [20].
Fig. 2 represents a schematic diagram of the micro-abrasivewear equipment (fixed ball equipment). In this device, a specificnormal load is applied to press a rotating ball (located betweentwo-coaxial shafts) into the surface of the coated specimen(placed in a pivoted specimen holder) in the presence of a specificabrasive slurry. The abrasive slurry was continuously agitated by
a magnetic stirrer and dropped steadily onto the rotating ball toproduce a spherical cup depression. In this investigation, highpurity silica from Lusosilica, Lda (Portugal) company, with anaverage particle size of 3.1 mm and hardness around 1000 HV wasused as abrasive to produce spherical cup depressions in the flatsurface of the coated sample. As shown in Fig. 3(a), the silicaparticles are highly irregular and angular. The particle sizedistribution is plotted in Fig. 3(b). The slurry was prepared withdifferent silica contents (25, 50 and 65 vol%) in distilled water. AnAISI 52100 steel ball bearing, 25.4 mm in diameter and with ahardness of approximately 740 HV, was used as the counterpartto produce wear scars in the sample surface. The specimen testedhas a hardness of approximately 225 HV. The rotation speed ofthe ball was kept at 75 rpm (0.1 m/s of tangential speed) in all thetests. The coating was tested using three different normal loadvalues (0.1, 0.2 and 0.5 N) and five test durations (50, 80, 150, 200and 300 rpm), respectively 4, 6.4, 12, 16 and 24 m of slidingdistance. The test conditions are summarized in Table 1.
In order to produce an evenly prepared surface, the specimenwas polished with SiC abrasive paper up to 1000 grit before weartests. A new ball was used in the abrasion tests. The ball’s surfacewas prepared using the run-in procedure described by Gee et al.[21], and the orientation of the ball was changed randomly beforeeach test. Following the tests, the wear scars were examined by ascanning electron microscope (SEM), in order to discriminate thedissimilar wear modes. Abrasion maps of the material volumeloss as function of the sliding distance were plotted for each valueof normal load used. The results were discussed in terms of eitherthe tribological properties or the dominant wear mechanismsinvolved. The material wear volume was calculated by using Eq. (1).In this equation b represents the crater chordal diameter of the
Fig. 4. Results of the micro-scale abrasion wear test for the different test
conditions used.
Table 1Test conditions.
Ball material Steel AISI 52100
Ball diameter (mm) 25.4
Abrasive Silica SS40 (mean size: 3.1 mm)
Abrasive concentration 25, 50 and 65 vol% in H2O
Test duration (rotations) 50, 80, 150, 200, 300
Normal load (N) 0.1, 0.2, 0.5
Speed (rpm) 75
F. Fernandes et al. / Wear 292–293 (2012) 151–158 153
spherical cup depression produced and R is the ball radius used. Theb value was calculated using the average of two perpendicularmeasurements performed at the wear scar chordal diameter, withthe help of a Mitutoyo Toolmaker Microscope with x–y micrometertable. All the data was then collected in a single graph, andrepresented as function of the volume loss, with the productbetween the sliding distance and the normal load, to identify theconditions that produce similar wear mechanisms. To quantify thespecific wear rate of the material a linear approach to Archard’s lawwas used. This law is given by Eq. (2). In this equation V representsthe wear volume, k is the specific wear rate, N is the normal appliedload and x is the sliding distance. Moreover, a statistical analysis wasapplied to estimate the error of measurements, as described else-where [22].
V ¼ ðf� b4Þ=ð64� RÞ ð1Þ
V ¼ k� N � x ð2Þ
Eq. (1) allows the approximate calculation of the wear volumeof a spherical cup depression, with an implied assumption thatthe wear scar has a spherical geometry conforming to that of theball bearing used. In order to elucidate users, an analysis of theerror of Eq. (1) is given in Appendix A. Moreover, a correctionfactor is proposed to enhance the accuracy of this equation. Theanalysis carried out in Appendix A, showed that for b/R values upto 0.35 the equation assures an error lower than 1%.
3. Results
The micro-scale abrasion results for the different test condi-tions are shown in Fig. 4. Each graph plots the volume loss (V) asfunction of the sliding distance (x), for each normal load. Theresults of the tests performed with the lowest load (see Fig. 4(a))reveal that there are no effective changes in the wear volume withthe change in the concentration of abrasive slurry, especially forthe scar produced with low sliding distance. Inversely, for thehighest load (see Fig. 4(c)), the results show that the materialvolume loss increases with increasing concentration of abrasiveand the difference is sharper for higher sliding distances. The testsconducted with an intermediary load (see Fig. 4(b)), displayedmixed behavior. The scars produced with the lowest concentra-tion of abrasive slurry showed lower wear volume than thoseproduced with higher concentrations of abrasive slurry (50% and65% in volume), which displayed similar values of materialvolume loss. Moreover, it is observed that the wear volume
F. Fernandes et al. / Wear 292–293 (2012) 151–158154
stabilizes with increasing sliding distance, whatever the concen-tration of abrasive used. This behavior could be associated withtwo effects: the running-in effect [23] and the variation of thecontact area/wear volume ratio throughout the duration of thetest. According to Blau [23], the running-in behavior is the netresult of simultaneous transitional processes occurring within theinterface, which may lead to non-linear evolution of the wear forshort sliding distances. Furthermore, at the start of the wear tests(with very small craters, consequently with small amounts of theradius of the spherical depression) it is well known that the areaof the crater grows much faster than the volume; at that point,the rate of pressure reduction is greater than the rate of change inthe volume, which may also justify the reduction in wear rate. Infact, the severity of contact index, proposed by Adachi andHutchings [6], is inversely proportional to the crater area; there-fore, a sudden increase in the area results in a reduction of theseverity index, which will induce a change in the wear rate. Thechanges in the wear volume described above are only partiallycorrelated with the test conditions, encouraging the investigationof the interactions between the abrasive slurry and the surface ofthe specimen in the contact zone. Scanning electron microscopy
Fig. 5. SEM morphologies of the worm surfaces of the sample performed with the com
and 4 m, (b) 0.5 N, 50%, 16 m and (c) 0.1 N, 25%, 16 m. Magnification of the wear mec
analyses were conducted on different spherical cup depressions inorder to identify the wear mechanism produced by the dissimilartest conditions.
Fig. 5 shows SEM micrographs of the dissimilar morphologiesof the spherical cup depressions induced by different test condi-tions of the micro-scale abrasion tests. As these images show, twodifferent wear mechanisms were identified, rolling abrasion (seeFig. 5(a)) and grooving abrasion (see Fig. 5(b)). Moreover, acombination of both wear mechanisms (see Fig. 5(c)) was alsoidentified for several contact conditions. In order to facilitateidentification of the wear mechanisms present in each wornsurface, magnified images are shown in Fig. 5(a1)–(c1). Rollingabrasion is characterized by multiple indentations in the wornsurface, (see Fig. 5(a1), whereas grooving abrasion (see Fig. 5(b1)is characterized by parallel grooves in the worn surface whichresults from the plastic deformation of the material. In turn, thesurface morphology, which displays a mixed-mode (rolling andgrooving abrasion), corresponds to a transition zone characterizedby a mixture of grooves and indentations, as shown in Fig. 5(c1).The transition of rolling to grooving abrasion occurs due to theembedding of particles in the surface. In this case grooving starts
bination of load, volume fraction of abrasive and sliding distance of: (a) 0.5 N, 50%
hanism present in the wear scar of: (a1) Figure a), (b1) Figure b), (c1) Figure c).
Fig. 6. (a) Scanning electron micrograph of the wear scar produced with 25% of
volume fraction of abrasive, load of 0.5 N and a sliding distance of 6.4 m: 1–ridging
wear mechanism zone, 2–grooving wear mechanism zone and 3–rolling wear
mechanism zone. (b) Magnification of the ridging wear mechanism zone.
Fig. 7. Wear modes observed in the wear scars produced by the different test
conditions.
Table 2Micro abrasion mechanism map as function of the
% of volume fraction of abrasives and loads used.
F. Fernandes et al. / Wear 292–293 (2012) 151–158 155
to occur in the zone of higher contact pressure, while rolling ismore commonly located in the surrounding area of less pressure,as shown in Fig. 5(c). Furthermore, in the tests performed with25 vol% abrasive and a load of 0.5 N, especially for the shortertests, the wear scars displayed areas with ridge wear mechanism,as Fig. 6 shows. Normally, this wear mechanism occurs for highvalues of normal load when the abrasive particles are absent fromthe contact region and contact is made directly between thesphere and the specimen tested, leading to ridging marks in thesurface [8], see zone 1 marked in Fig. 6(a). In this investigation theridging points will be disregarded since, in this type of wear, theabrasive particles do not influence the formation of the wear scarsand, therefore, the abrasion mechanism does not consider theeffect of the abrasive slurry.
% of volume fraction
of abrasiveLoad (N)
0.1 0.2 0.5
25 O // //
50 O y //
65 O y y
O–Rolling wear mechanism, y–Mixed-mode wear
mechanism and //–Grooving wear mechanism
4. Discussion
Considering the cumulative effect of the normal load multi-plied by the sliding distance (N� x) product, the wear map for thedifferent test conditions was plotted on a wear volume againstN� x graph by identifying the regimes of interactions based on
SEM observations. As can be seen in Fig. 7, three well definedzones (rolling abrasion zone, grooving abrasion zone and a mixed-mode wear) can be perceived. The graph reveals that differentcombinations of test conditions can produce the same wearmechanism. Moreover, it is observed that for some combinationsof concentration of abrasive slurry and load, the same wearmechanism is maintained even if the sliding distance is changed(conditions of (50 vol%; 0.1 N), (65 vol%; 0.1 N) and (25 vol%;0.2 N)). In these cases the wear volume always increases withincreasing sliding distance, even if the dominant wear mechanismis grooving or rolling abrasion. However, some conditions ofabrasive slurry concentration and load are more prone to induceone type of wear mechanism even when changing the slidingdistance. These changes are related to transitions in rolling tomixed-mode mechanism and mixed-mode to rolling wearmechanisms. For example, the wear scars produced with25 vol% abrasive using a load of 0.1 N produces both rollingabrasion and the mixed-mode wear mechanism depending onthe sliding distance. Table 2 summarizes the main wear mechan-isms produced by the different test conditions. The rolling,grooving and rolling–grooving abrasion regimes are differentiatedby using different symbols. The Table shows that low loads andhigh volume fraction of abrasive slurry enable three-body abra-sion, while grooving abrasion becomes stable at high loads andlow volume fraction of abrasive slurry. These last results are ingood agreement with previous studies [6,18].
Archard’s model is a suitable way to estimate a parameter ableto quantify the wear behavior of materials (the specific wearrate). However experimental results only fit the model well if asingle wear mechanism occurs in all the tests. The resultsdisplayed in Fig. 7 show that different combination of testconditions can produce the same wear mechanism. So, in thisstudy the experimental results were organized into two groupsdepending on the wear mechanism involved; rolling or grooving.
Fig. 8. (a) Results of the micro-scale abrasion wear test for the rolling and
grooving wear points and (b) error bars of the specific wear rate for a confidence
interval of 90%.
Table 3Results of the linearization analysis of the rolling and grooving zones.
Specific wear rate, k (mm3/N m)
Average STD Confidence interval 90% r2
Rolling zone 5.56�10�3 7.15�10�4 4.03–7.09�10�3 0.812
Grooving zone 7.79�10�4 8.21�10�5 6.20–9.39�10�4 0.937
F. Fernandes et al. / Wear 292–293 (2012) 151–158156
Subsequently, Archard’s model was applied to quantify thespecific wear volume characteristic of each wear mechanism.
The micro-scale abrasion test results of the rolling and groov-ing abrasion points are shown in Fig. 8, as is the material volumeloss function of the product of the sliding distance by the normalload. For each set of data, a straight linear trend was fitted to thepoints, obtaining the specific wear rate as the slope of therespective linear trend line, as shown by Eq. (2). A completereliability analysis of the specific wear rate is summarized inTable 3 and, for each case, the confidence bands are plotted inFig. 8(b) for a confidence level of 90%. The results show that thespecific wear rate obtained from the rolling experimental datapoints is around one order of magnitude higher than in thecorresponding grooving points. In fact, the rolling wear mode isa more effective wear mechanism than the grooving mode,leading to higher values of specific wear rate [5,18,24]. Further-more, the reliability of results indicates that the correlationcoefficient for the specific wear rate calculated from the rollingand grooving points is high, meaning that the quality of the linearfittings displayed in Fig. 8 is suitable. Therefore, according to
these results, the specific wear rate can be estimated using pointsproduced with different test conditions, since a single wear modeis sustained. Furthermore, these results lead to the conclusionthat the specific wear rate essentially depends on the wearmechanisms (rolling or grooving) involved and not on the testsconditions employed, since these do not produce changes in thewear mechanism.
5. Conclusion
(i)
In this investigation, silica slurry was used as an abrasive inmicro-scale abrasion equipment to study the influence ofdifferent test conditions on the dominant wear modes of aNi-based coating deposited by PTA on gray cast iron.
(ii)
The effect of different concentrations of abrasive slurry,sliding distances and normal loads on the micro-abrasion ofa Ni-based coating were investigated and showed significantdifferences between micro-abrasion rates and wear mechan-isms as a function of these parameters.
(iii)
The micro-abrasion mechanism maps were represented interms of volume loss as a function of the sliding distancemultiplied by the normal load. The results showed thatdifferent combinations of tests can produce the same wearmechanism. Low values of normal load and high contents ofabrasive particles enable rolling abrasion, while groovingabrasion is stabilized at high loads and low volume fractionof abrasive slurry.
(iv)
The micro-scale abrasion results involving rolling and groov-ing abrasion data, based on calculations of the specific wearrate, revealed that this parameter can be successfully calcu-lated using data points produced by dissimilar test condi-tions since they give rise to a single wear mode. This resultled to the conclusion that the specific wear rate essentiallydepends on the wear mechanisms involved and not onthe test conditions employed, since a single wear mode issustained (grooving or rolling abrasion).
(v)
Abrasion by rolling wear induced a specific wear rate aroundone order of magnitude higher than the value obtained undergrooving abrasion.
Acknowledgments
The authors wish to express their sincere thanks to thecompany ‘‘Intermolde’’, for supplying the coated samples, andthe Portuguese Foundation for the Science and Technology (FCT),through COMPETE program from QREN and to FEDER, for financialsupport in the aim of the project number ‘‘013545’’, as well as forthe grant (SFRH/BD/68740/2010).
Appendix A. Error analysis of the equation of the wearvolume
This section aims to elucidate authors about the resultant error byusing an approximate equation to calculate the volume of a sphericalcup depression induced by the micro-scale abrasion tests.
The original equation used to calculate the volume of thespherical cup depression is given by the Eq. (A1), where b and h
represent respectively the crater chordal diameter and the depthof the spherical cup depression (see Fig A.1). However, in micro-scale abrasion testing an approximate equation of Eq. (A1) iscurrently used (see Eq. (A2)). In this equation R represents the ballradius of the ball bearing used. The error of using an approximate
Fig. A.2. Error associated to the approximate equation to estimate the volume of a
wear scar without and with corrective factor..
Fig. A.1. Schematic diagram of a crater induced by the micro-scale abrasion tests..
F. Fernandes et al. / Wear 292–293 (2012) 151–158 157
equation can be calculated by using Eq. (A3).
V ¼ ð1=6Þ � f� h� ð3� ðb=2Þ2þh2Þ ðA1Þ
Vn¼ ðf� b4
Þ=ð64� RÞ ðA2Þ
error¼ ðVn�VÞ=V ðA3Þ
It is evident from Fig A.1 that increasing the depth h of thespherical cup depression increases the value of the chordaldiameter b of the wear scar, until a maximum value of b equalto 2R is reached. So, discretizing the value of b in small intervals itis possible to calculate the associated values of V and Vn for eachrelation between b and h. The value of h (given by Eq. (A4)), can beeasily deduced by subtracting the value of c from the value of R, asrepresented in Fig A.1. The value of c is obtained by applyingPythagoras’ Theorem at the triangle. Applying Eq. (A3) at thedissimilar combinations of b and h it is possible obtain the errorassociated by using an approximate equation to estimate thevolume of a spherical cup depression induced by the micro-scaleabrasion test. The results of the error are plotted in Fig A.2, dashedline, as function of the ratio between the crater’s chordal diameterand the radius of the ball with the relative error associated. As canbe seen in this figure, the resulting error is very small for lowratios of b/R, however with an increase in this ratio, the errorincreases. For example considering a spherical cup depressionwith a chordal diameter equal to the radius of the ball, the errorcommitted in the calculations is about 10 percent.
h¼ R�ðR2�ðb=2Þ2Þ1=2
ðA4Þ
The level of correlation between the ratios of the depth of thespherical cup depression and the ball radius with the error proved tobe linear. So, according to this relationship it is possible to establish a
correction factor for a desired value of error. For example, for adesired error less than 1% a correction factor can be calculated asindicated by Eq. (A5). By manipulating the equation it is possible todetermine the limits as a function of the ratio between b and R (seeEq. (A6)). Fig A.2 displays the error obtained when the approximateequation to estimate the volume of the wear scar is used, correctionfactor applied (continuous lines) and without the correction factor(dashed lines). As can be seen in the figure, the error of using theexpression with the corrective factor is one hundred times smallerthan that obtained with the expression without the corrective factor.Although, the correction factor gives more accurate results for lowratios between the chordal crater and the radius of the ball bearing,the approximate equation to evaluate the volume of a spherical cupdepression is appropriate, due to the small error induced.
[1] K. Bose, R.J.K. Wood, Optimum tests conditions for attaining uniform rollingabrasion in ball cratering tests on hard coatings, Wear 258 (2005) 322–332.
[2] A.J. Gant, M.G. Gee, A review of micro-scale abrasion testing, Journal ofPhysics D: Applied Physics 44 (2011) 073001.
[3] M.F.C. Andrade, R.P. Martinho, F.J.G. Silva, R.J.D. Alexandre, A.P.M. Baptista,Influence of the abrasive particles size in the micro-abrasion wear tests ofTiAlSiN thin coatings, Wear 267 (2009) 12–18.
[4] T.S. Eyre, Wear characteristics of metals, Tribology International 9 (1976)203–212.
[5] D. Braga, A. Ramalho, P.N. Silva, A. Cavaleiro, Study of abrasion resistance ofsteels by micro-scale tests, in: P.M. Vilarinho (Ed.), Advanced MaterialsForum III, Parts 1 and 2, Trans Tech Publications Ltd., Zurich-Uetikon, 2006,pp. 544–548.
[6] K. Adachi, I.M. Hutchings, Wear-mode mapping for the micro-scale abrasiontest, Wear 255 (2003) 23–29.
[7] R.C. Cozza, D.K. Tanaka, R.M. Souza, Friction coefficient and wear modetransition in micro-scale abrasion tests, Tribology International 44 (2011)1878–1889.
[8] M.M. Stack, M. Mathew, Micro-abrasion transitions of metallic materials,Wear 255 (2003) 14–22.
[9] R.J.K. Wood, Tribo-corrosion of coatings: a review, Journal of Physics D:Applied Physics 40 (2007) 5502–5521.
[10] R.C. Cozza, J.D.B. de Mello, D.K. Tanaka, R.M. Souza, Relationship between testseverity and wear mode transition in micro-abrasive wear tests, Wear 263(2007) 111–116.
[11] E. Rabinowicz, L.A. Dunn, P.G. Russell, A study of abrasive wear under three-body conditions, Wear 4 (1961) 345–355.
[12] J.C.A. Batista, A. Matthews, C. Godoy, Micro-abrasive wear of PVD duplex andsingle-layered coatings, Surface and Coatings Technology 142–144 (2001)1137–1143.
[13] J.A. Williams, A.M. Hyncica, Mechanisms of abrasive wear in lubricatedcontacts, Wear 152 (1992) 57–74.
[14] J.A. Williams, A.M. Hyncica, Abrasive wear in lubricated contacts, Journal ofPhysics D: Applied Physics 25 (1992) A81–A90.
[15] F.J.G. Silva, R.B. Casais, R.P. Martinho, A.P.M. Baptista, Role of abrasivematerial on micro-abrasion wear tests, Wear 271 (2011) 2632–2639.
[16] R.I. Trezona, D.N. Allsopp, I.M. Hutchings, Transitions between two-body andthree-body abrasive wear: influence of test conditions in the microscaleabrasive wear test, Wear 225–229 (1) (1999) 205–214.
[17] M.M. Stack, M.T. Mathew, Mapping the micro-abrasion resistance of WC/Cobased coatings in aqueous conditions, Surface and Coatings Technology 183(2004) 337–346.
[18] A. Ramalho, J.R. Perez, A.R. Garcia, Microabrasion resistance of nanostruc-tured plasma-sprayed coatings, International Journal of Surface Science andEngineering 5 (2011) 250–260.
[19] P.H. Shipway, J.J. Hogg, Wear of bulk ceramics in micro-scale abrasion—therole of abrasive shape and hardness and its relevance to testing of ceramiccoatings, Wear 263 (2007) 887–895.
[20] F. Fernandes, B. Lopes, A. Cavaleiro, A. Ramalho, A. Loureiro, Effect of arccurrent on microstructure and wear characteristics of a Ni-based coatingdeposited by PTA on gray cast iron, Surface and Coatings Technology 205(2011) 4094–4106.
F. Fernandes et al. / Wear 292–293 (2012) 151–158158
[21] M.G. Gee, A.J. Gant, I.M. Hutchings, Y. Kusano, K. Schiffman, K.V. Acker,S. Poulat, Y. Gachon, J.v. Stebut, P. Hatto, G. Plint, Results from an inter-laboratory exercise to validate the micro-scale abrasion test, Wear 259(2005) 27–35.
[22] A. Ramalho, A reliability model for friction and wear experimental data, Wear269 (2010) 213–223.
[23] P.J. Blau, On the nature of running-in, Tribology International 38 (2005)1007–1012.
[24] K. Adachi, I.M. Hutchings, Sensitivity of wear rates in the micro-scale
abrasion test to test conditions and material hardness, Wear 258 (2005)318–321.
Annex E
123 Filipe Daniel Fernandes
Annex E
F. Fernandes, T. Polcar, A. Loureiro, A. Cavaleiro, Room and high temperature
tribological behavior of Ni-based coatings deposited by PTA on gray cast iron,
(2014), under review, “Tribology International”.
Annex E
125 Filipe Daniel Fernandes
Room and high temperature tribological behavior of Ni-based coatings deposited by
PTA on gray cast iron
F. Fernandes1,*
,T. Polcar2,3
, A. Loureiro1, A. Cavaleiro
1
1CEMUC - Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis
Santos, 3030-788 Coimbra, Portugal.
2Department of Control Engineering Czech Technical University in Prague Technicka 2,
Prague 6, 166 27 Czech Republic.
3n-CATS University of Southampton Highfield Campus SO17 1BJ Southampton, UK.
[15] Kesavan D, Kamaraj M. The microstructure and high temperature wear performance of a
nickel base hardfaced coating. Surf Coat Technol. 2010;204:4034-43.
[16] Mironova-Ulmane N, Kuzmin A, Steins I, Grabis J, Sildos I, Pärs M. Raman scattering
in nanosized nickel oxide NiO. Journal of Physics: Conference Series. 2007;93:012039.
[17] Jian S-R, Chen Y-T, Wang C-F, Wen H-C, Chiu W-M, Yang C-S. The Influences of H2
Plasma Pretreatment on the Growth of Vertically Aligned Carbon Nanotubes by Microwave
Plasma Chemical Vapor Deposition. Nanoscale Research Letters. 2008;3:230-5.
Annex E
144 Filipe Daniel Fernandes
[18] Pandya SN, Nath SK, Chaudhary GP. Friction and Wear Characteristics of TIG
Processed Surface Modified Gray Cast Iron2009.
[19] de Faria DLA, Venâncio Silva S, de Oliveira MT. Raman microspectroscopy of some
iron oxides and oxyhydroxides. Journal of Raman Spectroscopy. 1997;28:873-8.
[20] Ferrari AC, Robertson J. Raman spectroscopy of amorphous, nanostructured, diamond–
like carbon, and nanodiamond. Philosophical Transactions of the Royal Society of London
Series A: Mathematical, Physical and Engineering Sciences. 2004;362:2477-512.
[21] Li YQ, Qiu T. Oxidation behavior of boron carbide powder. Materials Science and
Engineering: A. 2007;444:184-91.
[22] Lavrenko VA, Gogotsi YG. Influence of oxidation on the composition and structure of
the surface layer of hot-pressed boron carbide. Oxid Met. 1988;29:193-202.
Annex F
145 Filipe Daniel Fernandes
Annex F
F. Fernandes, A. Ramalho, A. Loureiro, J.M. Guilemany, M. Torrell, A.
Cavaleiro, Influence of nanostructured ZrO2 additions on the wear resistance of Ni-
based alloy coatings deposited by APS process, Wear, 303 (2013) 591-601.
Wear 303 (2013) 591–601
Contents lists available at SciVerse ScienceDirect
Wear
0043-16http://d
n CorrE-m
journal homepage: www.elsevier.com/locate/wear
Influence of nanostructured ZrO2 additions on the wear resistanceof Ni-based alloy coatings deposited by APS process
F. Fernandes a,n, A. Ramalho a, A. Loureiro a, J.M. Guilemany b, M. Torrell b, A. Cavaleiro a
a CEMUC—Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, Portugalb Thermal Spray Centre (CPT-UB), University of Barcelona, C/Martí i Franques n 1, Barcelona, Spain
a r t i c l e i n f o
Article history:Received 19 December 2012Received in revised form8 April 2013Accepted 15 April 2013Available online 22 April 2013
48/$ - see front matter & 2013 Elsevier B.V. Ax.doi.org/10.1016/j.wear.2013.04.012
esponding author. Tel.: +351 239 790 745; faxail address: [email protected] (F. Fer
a b s t r a c t
In the present investigation, the influence of the addition of nanostructure zirconia particles on themicrostructure, micro-hardness and wear performance of a Ni-based alloy (Colmonoy 88) deposited byatmospheric plasma spraying (APS) on low carbon steel has been reported. Two different procedureswere tested: (i) spraying powders of Colmonoy 88 and zirconia mixed by mechanical alloying and (ii)spraying powders separately using a dual powder injection system available at the APS equipment. Themicrostructure and the mechanical properties of coatings were characterized by scanning electronmicroscopy/energy dispersive X-ray analysis (SEM-EDS), X-ray diffraction (XRD) and micro-hardnessmeasurements. The tribological properties were evaluated at room temperature in reciprocating slidingwear equipment. The results indicate that the as-sprayed modified coatings were mainly composed byNi, Ni–Cr–Fe, Cr23C6, Cr5B3, and tetragonal zirconia. Evenly distribution of zirconia can be seen in thecoatings produced by powders prepared by mechanical alloying, while dispersive ones can be seen in theother case. Hardness and wear resistance of coatings is increased with nanostructured zirconia additions,while their friction coefficient is decreased. Coatings produced with mechanical alloying show thehighest wear resistance of all tested coatings. Nanostructured ZrO2 coating displays the worst wearresistance.
& 2013 Elsevier B.V. All rights reserved.
1. Introduction
Improvement of the surface properties by thermal sprayinghard and wear resistant materials is a commonly used industrialpractice [1–3]. The base material provides the overall mechanicalstrength of the components while coatings provide a way ofextending the limits of their use at the upper end of theirperformance capabilities.
A wide variety of coating materials such as cobalt, iron andnickel alloys can be successfully applied by thermal sprayingtechnologies to protect the surface of components subjected toharsh environments [4–6]. Superalloys and cermet APS (atmo-spheric plasma spraying) coatings have been widely employed toimprove the oxidation, corrosion, abrasion and wear resistance ofengineering components, such as plungers, molds, wearing plats,turbines, tools, etc, whose surface is submitted to extreme tribo-logical conditions in service [7–10]. Nickel-based alloys have beenespecially used in the protection of parts due to their uniquecombination of properties (mechanical, tribological, and hightemperature properties) [11–14]. The microstructure and the wear
ll rights reserved.
: +351 239 790 701.nandes).
behavior of cobalt, iron and nickel alloys coatings have beenstudied using several alloy compositions, several processes ofdeposition and different substrates. Regarding the wear resistanceof these coatings, some special composite systems were studiedand developed. The incorporation of hard and stable carbides andoxide phases, such as CrC, SiC, WC, Al2O3, TiO2, ZrO2, CeO2, etc, inthese coatings, has been reported as a way to improve their wearresistance. For example, Hou et al. [15] showed the beneficialeffect of nano-Al2O3 particles on the microstructure and wearresistance of a nickel-based alloy coating deposited by plasmatransferred arc (PTA). Harsha et al. [1] studied the influence of CrCon the microstructure, microhardness and wear resistance of anickel-based alloy deposited by flame spraying process. Theyobserved that the wear resistance of the Ni modified coating isincreased relatively to the CrC-free Ni-based alloy. Wang et al. [16]investigated the effect of three nano-particles additions (Al2O3,SiC, CeO2) on the high temperature wear behavior of a Ni-basedalloy coatings produced by laser cladding technique and reportedthat the addition of these nanoparticles increased the wearresistance of the coatings. Regarding the effect of incorporationof ZrO2 particles in a Ni matrix, it has also been extensively studied[17–20]; however, only electrodeposited coatings have been con-cerned. Despite of these studies to improve the wear resistance ofNi electrodeposited coatings, at our knowlodgment no reports
have been published as regards to the effect of nanostructuredzirconia additions on the properties of Ni-based thermal sprayedcoatings. Since zirconia is a ceramic material of high level of
Table 1Nominal chemical composition (wt.%) of the as-received powders.
Powders W B C Cr Fe Si Ni
Colmonoy 88 13.5 2.6 0.6 14.1 3.5 3.7 Balance
HfO2 Y2O3 ZrO2
Nanostructured ZrO2 2.5 7.5 Balance
Table 2Plasma spraying parameters of the dissimilar coatings.
hardness, wear and oxidation resistance, its use as reinforcementmaterial in nickel-based alloy coatings deposited by thermalspraying processes should be investigated.
Therefore, the main goal of the present work was to study theeffect of nanostructured zirconia additions on the microstructure,micro-hardness and wear performance of a nickel-based hardfacingalloy deposited on low carbon steel by atmospheric plasma spraying(APS). The coatings with nanostructured zirconia additions wereproduced by spraying either powders prepared by mechanical alloy-ing or separate powders using a dual system of powder injection.The microstructure, the mechanical and tribological propertieswere characterized by scanning electron microscopy/energy disper-sive X-ray analysis (SEM-EDAX), X-ray diffraction (XRD), Vickersindentation tests and reciprocating sliding wear tests. All the
powders, (c) powders prepared by mechanical alloying, (condition with 40% ZrO2).
F. Fernandes et al. / Wear 303 (2013) 591–601 593
properties of the doped coatings were compared with the un-modified nickel and nanostructured zirconia coatings.
2. Experimental procedure
2.1. Materials
In this investigation commercial powders of a Ni-based alloy(Colmonoy 88) and nanostructured zirconia from the “Wall Colmonoy”and “Innovnano” companies, respectively, were used as feedstock toproduce coatings with Atmospheric Plasma Spraying process (APS)onto a low carbon steel substrate (AISI 8620). The Ni-based alloypowders have been characterized as a spherical morphology with anaverage size D50 of 45 mm and density of 7900 (kg/m3), whilstnanostructured ZrO2 powders have an average primary particle sizeof 60 nm with granule size D50 of 60 mm and density 1700 (kg/m3).The chemical compositions of the as-received powders are displayedin Table 1. The plasma sprayed coatings were deposited with an APSA3000 system from “Sulzer Metco”, using either pure as-receivedpowders and a dual system for powder injection available at theAPS equipment (designated as “Dual”) or mixtures of powdersprepared by mechanical alloying (MA) process (designated as “MA”).The substrate material was grit-blasted with alumina grade 24 from“Alodur Germany” before coating (which induced a correspondingsuperficial roughness (Ra) of 5 mm), to increase the surface roughnessand achieve the proper mechanical interlocking between the coatingand the substrate. For pure colmonoy and zirconia coatings and Dualprocedure, the substrates were fixed positioned, being the spraying ofthe surface ensured by the movement of the torch. On the other hand,
Fig. 2. X-ray diffraction pattern of the dissimilar powders used in the depositions.
Fig. 3. Energy dispersive spectroscopy analysis (EDS) of a powde
in the MA method the substrates were placed in a rotatory table andthe spraying was performed by combining this rotation with thevertical movement of the torch. Before final depositions, preliminarycoatings were produced changing the main deposition parametersand, then, characterized in order to properly select the depositionsconditions that allowed achieving the best results. Moreover, duringdepositions the substrates were cooled down by two pressurized airguns keeping the temperature below 150 1C, in order to avoid crackingand residual stresses formation.
Powders mixtures were prepared by mechanical alloying in aFritsch planetary ball mill using a 250ml hardened steel vial andfifteen balls with 20 mm diameter of the same material. Two propor-tions of nanostructured zirconia were added to the nickel-based alloypowder, 20% and 40% in volume, i.e. 5.1 and 12.5 gr of nanostructuredZrO2 powder for 94.9 and 87.4 gr of nickel-based alloy powder,respectively. The MA process was carried out at 400 rpm for 2 h in aprotective atmosphere of argon. After 1 h of milling, the process wasinterrupted for 10 min to cool the vial and to reserve rotation. Finally,after milling, the powder was mechanical sieved to obtain particleswith a size distribution in the range [28–71 mm]. Two differentcoatings also containing nanostructured zirconia additions of 20%and 40% were also performed by using the dual system of powderinjection available at the APS equipment. In this case the proportionwas easily achieved by controlling the flow of powders through therotation control of the powder feeders. Henceforth, coatings producedusing as received pure powder of nickel-based alloy and nanostruc-tured zirconia will be designated as “Ni”, and “ZrO2”, respectively,coatings produced with powders prepared by mechanical alloyingusing 20% and 40% of nanostructured ZrO2 will be termed as “Ni+20%MA” and “Ni+40% MA”, respectively, and coatings deposited using thedual system of powder injection with 20% and 40% of nanostructuredZrO2 will be designated as “Ni+20% Dual” and “Ni+40% Dual”,respectively. The process parameters employed in the dissimilarcoating depositions are shown in Table 2.
2.2. Powders and coatings characterization
Powders morphology were observed and characterized by scan-ning electron microscopy with X-ray spectroscopy (SEM/EDS) andX-ray diffraction (XRD) using Co Kα radiation. Transverse sectionof APS coatings was polished using conventional metallographicprocedure which consisted grinding followed by polishing. Themicrostructure of coatings was studied and characterized by opticalmicroscope (OM) and SEM/EDS. Further, the structure of the coatingswas studied by X-ray diffraction. The micro-hardness of coatings wasdetermined by Vickers testing using a 3 N load at 15 s holding time. Atotal of 30 measurements of hardness were done in each coating.
r produced by mechanical alloying (condition Ni+40% ZrO2).
F. Fernandes et al. / Wear 303 (2013) 591–601594
2.3. Wear tests
Wear behavior of APS coatings was studied using reciprocatingwear equipment. An harmonic wave generated by an eccentric androd mechanism imposed a stroke length of 2.05 mm at a frequencyof 1 Hz. A detailed description of the equipment can be found inreference [21]. Before wear tests, the surface of coatings weregrounded and then polished using a 3 mmdiamond past. A superficialroughness Ra of: 0.51, 0.65, 0.49 and 0.98 mm was measuredrespectively for Ni, Ni Dual, Ni MA and ZrO2 coatings. In the scopeof the present study, the values of normal load applied to the coatedsamples were 5, 7, 8.5 and 10 N. All the tests were conducted at aconstant rotational speed of 190 rpm during 2 h. The acquisition time
Fig. 4. SEM morphology of the: (a) Ni-based alloy coating, (b) nanostructured ZrO2 coatand Ni+40% MA.
of signal used was 60 s. A soda-lime glass sphere with 500 HV0.5 ofhardness and 10 mm in diameter was used as counterpart, in orderto study the interaction of glass with the different coatings. Afterwear tests, the volume loss was estimated through the transverseprofile of each wear scar at the middle and near the end of both sidesof the scar. To ensure the reproducibility of results, a set of three scarsfor each test condition was performed at the coating, and the volumeloss calculated by using the volume average of these marks. Anapproach of the Archard's law was applied; see Eq. (1), in order toobtain the specific wear rate. In this equation, V represents the wearvolume, P is the normal load, k is the specific wear rate and l thesliding distance. To estimate the error of the measurements, astatistical analysis described elsewhere by A. Ramalho [22] was used.
ing, (c) and (d) Ni+20% Dual and Ni+40% Dual, respectively, (e) and (f) Ni+20% MA
F. Fernandes et al. / Wear 303 (2013) 591–601 595
After wear tests, the morphology of depressions was observed andcharacterized by scanning electron microscopy.
V ¼ k� P � l ð1Þ
3. Results and discussion
3.1. Powders characterization
Fig. 1(a–c) shows SEM morphologies of nickel-based alloy,nanostructured zirconia and one of the MA prepared powders.
Fig. 5. SEM-EDS analysis performed at the: (a) Ni coating, (b) Ni+40% Dual coating, (c) N
Ni and ZrO2 powders clearly display spherical-shaped particles, whileMA powders have irregular shapes. XRD patterns in Fig. 2 shows thatNi powders consist essentially in a Ni solid solution (ICDD card 01-1258) with some traces of chromium carbide (ICDD card 85-1281)and tungsten (ICDD card 47-1319). This indexation is in agreementwith the results from the literature where similar phases weredetected [23]. Nanostructured ZrO2 powders are essentially formedby tetragonal zirconia. In respect to MA powders XRD pattern revealthe same phases indexed from the pure Ni and nanostructured ZrO2
powders, suggesting only physical mixing, not having occurredchemical reactions. Fig. 3 shows a SEM-EDS analysis performed at a
i+40% MA coating. SEM-EDS spectra of: (d) point 1, (e) zone 2, (f) point 3, (g) zone 4.
Fig. 6. X-ray diffraction pattern of the dissimilar coatings.
Fig. 7. Microhardness of coatings.
Fig. 8. Volume loss of: (a) coatings, (b) ball; after wear tests as function of theapplied load.
Fig. 9. Evolution of the wear behavior as a function of the parameter “normal load� sliding distance”.
Table 3Results of the linearization analysis of the dissimilar coatings.
CoatingSpecific wear rate, k(mm3/N�m)
Average STD confidence interval 90% r2
Ni 1.68�10−5 3.06�10−6 7.83�10−6–2.57�10−5 0.938Ni+20% Dual 4.45�10−6 1.14�10−6 1.12�10−6–7.79�10−6 0.884Ni+40% Dual 5.26�10−6 8.65�10−7 2.74�10−6–7.78�10−6 0.949Ni+20% MA 3.83�10−6 5.56�10−7 2.21�10−6–5.46�10−6 0.960Ni+40% MA 1.90�10−6 3.97�10−7 7.38�10−7–3.06�10−6 0.919ZrO2 3.03�10−4 7.15�10−5 9.48�10−5–5.12�10−4 0.900
F. Fernandes et al. / Wear 303 (2013) 591–601596
MA powder (Ni+40% ZrO2). As the EDS spectra displays, the powderis essentially formed by nickel, zirconia, oxygen and tungsten,supporting the results from the XRD analysis.
3.2. Microstructure
Fig. 4 shows SEM micrographs of the cross section of thecoatings. All coatings present the typical morphology of plasmaspraying, with pores, lamellae, partially melted and un-meltedparticles and good adhesion to the substrate. ZrO2 (see Fig. 4b) andNi+ZrO2 Dual coatings (Fig. 4(c and d)) show a higher level ofporosity than the other coatings. Very often, small pores werefound within the flattened particles and result from shrinkageporosity and big pores were normally located between lamellaeand were caused by gas porosity phenomenon [24]. In all cases themelting states of particles influence the level of porosity and,usually, low melting state of particles gives rise to a porousmicrostructure. Furthermore, pure ZrO2 coating, (see Fig. 4b)reveals the presence of some cracks in the coating, resulting fromtensile residual stresses generated during cooling down to roomtemperature. Ni and MA coatings (see Fig. 4(a), (e) and (f)) displaya compact and homogenous structure, while Dual coatingsrevealed a microstructure full of semi-melted particles, suggestinga brittle character based on low level of agglomeration betweenparticles. It is well known that in the thermal spraying processesthe melting level depends on the thermal energy added to theparticles, on the plasma temperature and on the particles flyingtime. During plasma spraying, molten particles impact the surface
F. Fernandes et al. / Wear 303 (2013) 591–601 597
of substrate, then deform, solidify and transform into lamellae.During spraying, the particles can be in all of the followingstates on impact: fully molten, superheated, semi-molten andmolten then re-solidified [24,25], depending on the depositionparameters. It is well known that in plasma spraying the electricarc current, the primary plasma gas flow rate, the second plasmagas flow rate and powder size are the main parameters thatinfluence the in-flight particle behaviors. From Table 2, comparingthe process parameters for the Ni and Ni+20% and 40% ZrO2 Dualcoatings, differences can be observed in relation to the spraydistance, primary and secondary gas flow rates, and powder feedrate. According to the literature, an increase of argon flow rateleads to a gentle decrease of particle temperature but a rapidincrease of velocity [26]. On the other hand, increasing thehydrogen flow rate increases both, temperature and velocity,though temperature is more sensitive to the hydrogen flow ratethan velocity. So, due to the lower primary gas flow rate (Ar) andhigher secondary gas flow rate (H2) used in the Ni +ZrO2 Dualcoatings, would be expected a great level of melting particles andconsequently an homogenous structure, if the particles do notoxidize during the in-flight time. However, such fact was notobserved. The lower primary gas flow rate (Ar) used in theproduction of these coatings might suggest that independentlyof the high melting state of particles, the velocity given by the flowcould be not enough to accelerate the particles onto the specimensurface. Nevertheless, the same process parameters were used inthe deposition of the Ni+20% MA coating, with exception of thespraying distance, and the microstructure revealed to be homo-genous and compact. In fact, the spraying distance also plays animportant role in the thermal treatment of particles during theirin-flight time [27]. Thereby, for short spraying distances, due tothe short exposure of particles in the flame, the energy added bythe flame could be not enough to semi-molten or molten theparticles, giving rise to a microstructure similar to those observedfor these coatings. Furthermore, it can also be speculated that thetwo different ways to add zirconia at the nickel alloy can changethe entropy of the system, and therefore change the running-on of
Fig. 10. Variation of the friction force of coatings with applied load for: (a) Ni coating,
the deposition. In fact, according to the literature the nature ofpowders has a detrimental effect in their interaction with theplasma spraying flame and, therefore, on the properties of thecoatings [28]. The low level of melting particles can also explainthe high level of porosity observed in these coatings.
For better characterization and perception of the differentcoatings microstructure, magnifications of the Ni, Ni+40% Dualand Ni+40% MA coatings, as well as EDS analyses are shown inFig. 5. The EDS analysis revealed that Ni coating is essentiallyformed by Ni, W, Cr and Fe, which matches well with the chemicalcomposition of the as received Ni powders. Moreover, smallparticles rich in W (white phase) can be observed evenly dis-tributed in the microstructure. The semi-melted particles in themicrostructure of the Ni+40% Dual coating are from the samenature of the Ni coating matrix, therefore, they can be identified assemi-melted Colmonoy powders. Similar to Ni coating a whitephase rich in W is also detected in the microstructure, co-existingwith a coarse gray phase, entrapped between the boundaries ofthe semi-melted particles, revealed by EDS to be zirconia. Such adistribution is in the basis of a brittle behavior. Contrary to thiscoating, Ni+40% MA displays a compact and homogeneous micro-structure, with small zirconia particles evenly distributed alongthe microstructure improving the coating toughness. Therefore,from these results, it can be inferred that the addition of nano-structured zirconia can be successfully achieved by using bothprocedures, powders prepared by mechanical alloying or dualprojection. However, in coatings deposited with MA powderszirconia is finer and more homogeneously distributed in theNi-matrix giving rise to better mechanical performance.
XRD patterns of the coatings are displayed in Fig. 6. As thespectrum reveals, the nickel-based coating consists mainly of Ni,Ni-Cr-Fe, Cr23C6 and Cr5B3 phases. These results are in accordancewith the literature where similar phases were identified [23].Moreover, all the coatings with nanostructured zirconia additionsshowed diffraction lines corresponding to the same phases identi-fied at both the Ni and ZrO2 coatings. In the latter, is mainlyformed by tetragonal zirconia.
Fig. 7 shows that nanostructured ZrO2 additions increases thehardness of the nickel-based alloy. The same behavior was reportedby other authors after addition of hard particles such as (Al2O3, CeO2,SiC, WC) at a nickel alloy [1,3,16]. However, MA coatings have higherhardness than Dual ones. This result is coherent with either thelower porosity or the finer and more homogeneous distribution ofnanostructured zirconia in Ni+ZrO2 MA coatings.
3.4. Wear behavior
The volume loss due to wear of the coatings and counterpartincreases with increasing applied load, as it is shown in Fig. 8(a) and(b), respectively, being this trendmore notorious for ZrO2 coating. As itwill be shown later, the higher volume loss for this coating can becorrelated with the brittle behavior as displayed on its worn surface.Nanostructured zirconia additions have a positive effect, decreasingthe volume loss of material, independently of the deposition proce-dure used. However, efficiency seems to be higher in Ni+ZrO2 MAcoatings which display higher decreases of the volume loss than Ni+ZrO2 Dual coatings. Furthermore, increasing zirconia content in MAcoatings continues to have a beneficial effect whereas the inverse isobserved in Dual coatings. The better wear performance of Ni+ZrO2
MA coatings can be attributed to the combination of a higherhardness, an evenly distribution of smaller zirconia domains in theNi-based matrix and a more compact microstructure. In fact, thepresence of hard phases increases the overall hardness of the coatingand, on the other hand, smaller and uniformly distributed hard
Fig. 11. SEM morphologies of the worn surfaces tested at 8.5 N load of the:
domains in a soft matrix gives rise to tougher materials, whichcombined lead to a decrease of the volume loss due to wear. Similarbehavior was observed by other author after adding hard particles to anickel based alloy [1,15]. In the case of Ni+ZrO2 Dual coatingsnanostructured hard ZrO2 phases have initially a similar influence,by increasing the hardness. However, the agglomeration of ZrO2
particles, localized in-between Ni-based lamellae makes microstruc-ture less tough with its consequent detrimental effect on the wearperformance when high contents of the hard phase exist.
In order to normalize the wear results, they were plotted in Fig. 9in terms of material volume loss as function of the product of slidingdistance by the applied load. For each set of data, a straight lineartrend was fitted in order to obtain the specific wear rate from theslope (see Eq. 1). Further, a complete reliability analysis of the specificwear rate is presented in Table 3 for a confidence interval of 90%.Because of the high volume loss displayed by the ZrO2 coating, onlythe point produced with lower load and the beginning of the straightadjustment were plotted in Fig. 9. Crossing the information of Fig. 9and Table 3 it is possible conclude the specific wear rate ofNi-modified coatings is lower than that of the un-modified Ni coating.Ni+40% ZrO2 dual coating display the lowest specific wear rate, beingthis result attributed to the higher hardness displayed by this coating.Further, coatings produced by the dual system of powder injectiondisplay higher specific wear rate than coatings produced by powdersprepared by mechanical alloying. ZrO2 coating displays the highestspecific wear rate, being one and two orders of magnitude greaterthan the un-modified and modified coatings, respectively.
The average values of the frictional force of the Ni, Ni+40% Dual,Ni+40% MA and ZrO2 coatings linearly increases with the applied
load, as shown in Fig. 10. This is in good agreement with Amontons–Coulomb model. By fitting a linear trend to the experimentalresults, the friction coefficient calculated from the slope decreases
Fig. 12. Magnification of the worn surfaces tested with a load of 8.5 N of the: (a) Ni(e) point A, (f) point B.
with the nanostructured zirconia additions, independently of thedeposition method. Pure nanostructured zirconia shows higherfriction coefficient than composite coatings, but lower than the
pure nickel one. The changes in the specific wear rate and frictioncoefficient described above are only partially correlated with theapplied test conditions and properties of the materials, which canproduce dissimilar wear mechanisms during wear tests, and there-fore dissimilar volume loss of coatings. These reasons have encour-aged the investigation of the interaction between the pairspecimens-counterpart. Scanning electron microscopy (SEM) ana-lysis was conducted on the dissimilar worn surfaces of coatings inorder to identify the main wear mechanisms produced.
Typical SEM morphologies of worn surfaces of the coatingstested with 8.5 N against a glass sphere are shown in Fig. 11.Distinct wear mechanisms were identified in the worn surfaces. InNi and Ni+ZrO2 Dual coatings significant amount of adherentmaterial is shown in the wear scar, while Ni+ZrO2 MA and ZrO2
coatings show clean worn surfaces. EDS analysis performed inadhered debris, marked in Fig. 12(a) and (b) by letters A and B,allows concluding that they are made of silicon oxides and Ni,suggesting a mechanical mixing by the friction process, joining thesofter Ni phase with the silica type wear debris from the ball wear.Being softer, the compact wear debris can stick at the worn trackas well as at the ball wear scar, as can be observed in Fig. 13.During contact the friction generated by the movement of thecounterpart against the surface specimen, causes an increase oftemperature promoting local plastic deformation. At this time,both materials will adhere to each other. In the wear track of theNi coating, see Fig. 12(a), the wear debris tend to be incorporatedin the specimens, more precisely in the boundaries of the semi-melted powders, remaining adherent and leading to removal ofmaterial. This mechanism explains simultaneously the high fric-tion values reported for Ni coating. On the other hand, ZrO2
coating shows clean wear scars without traces of adhered material.The surface has a brittle appearance, where micro cracks can beseen, as shown in Fig. 12(d), revealing a rough surface from whichparticles of several tens of micrometers were detached. Thismorphology suggests mechanical interlocking during the contactwhich can explain the higher values of the friction coefficient.Furthermore, it should promote a strong abrasion in the counter-part, giving rise to much higher wear rate as shown in Fig. 8(b).Moreover, the evolution of the wear amount with the increase ofnormal load, (see Fig. 8), allow identifying a made severe wearvolume for loads higher than 7 N, which should be correlated to atransition for a made brittle behavior. Similar behavior wasobserved by Chen et al. [29] during the evaluation ofun-lubricated wear properties of plasma-sprayed nanostructuresand conventional zirconia coatings. The wear scars of Ni+ZrO2
Fig. 13. Optical micrograph of the ball wear scars tested against the Ni coating with7 N of load.
Dual coatings show a mix morphology of pure Ni and ZrO2
coatings, i.e. adhered debris can be observed over a roughbrittle-appearance surface, whose number decreases with nanos-tructured zirconia additions. In these coatings, zirconia particlesare localized in the boundaries of the semi-melted Ni-basedpowders (see Fig. 4c and d). Despite of the particles operating asreinforcement of the coatings, leading to a lower volume loss, thearea of the nickel exposed to the glass sphere remains higher.Therefore, it is expected similar debris adhesion as for Ni coating.At the same time, due to the continuous removal of material, ZrO2
particles in the boundaries of the semi-melted powders loses theirsupport capacity being removed from the boundaries. Theseparticles will slide through the wear track acting as a removal ofthe wear debris. Increasing the volume of nanostructured zirconiafrom 20 to 40%, higher levels of brittleness in the boundaries areachieved, increasing the volume loss of material, as reportedbefore, and giving rise to the rough and brittle appearance shownbelow the adhered debris. The higher level of porosity of thesecoatings also accounts for higher levels of volume loss of material.
The coatings produced from MA powders show also clean wearsurfaces but with a different failure type from pure zirconia coating. Ni+ZrO2 MA coatings reveal surfaces with scratches identifying somesort of grooving abrasion, see Fig. 12(c). The uniform distribution ofsmall nanostructured zirconia particles and the higher hardness andtoughness of these coatings avoids the plastic deformation, theadhesion and the liberation of large wear debris either from thecoating or the counterpart. The smooth worn surfaces can be in thebasis of the low friction coefficient [30]. In these coatings, the matrixcutting occurs through a process of two body abrasion, as shownFig. 12(c). Despite of the dissimilar wear mechanisms observed in thewear track of the coatings, during the wear tests it was observed thatmost of the material resulting from the wear was blown out, due tothe movement of the ball.
4. Conclusions
�
The influence of nanostructured zirconia additions on themicrostructure, micro-hardness and wear performance of anickel-based hardfacing alloy deposited by atmosphericplasma spraying (APS) on low carbon steel is reported in thepresent paper.
�
The coatings containing nanostructured zirconia additionswere produced either using powders prepared by mechanicalalloying or using separate powders that are sprayed simulta-neously via a dual system of powder injection.
�
Nanostructured zirconia can be successfully incorporated intonickel-based alloy powders by mechanical alloying. Coatingsproduced with these powders present evenly distributed zir-conia as opposed to coatings produced by separate powderinjection, where the nanostructured zirconia particles wereentrapped in the boundaries of the semi-melted Ni lamellae.
�
The microstructure of the nickel coating consists mainly of Ni,Ni–Cr–Fe, Cr23C6 and Cr5B3 phases. The same diffraction lineswere identified in the modified nickel coatings with tetragonalzirconia, independently of the procedure used in depositions.
�
The hardness and wear behavior of coatings was improvedwith nanostructured zirconia additions while the frictioncoefficient is decreased.
�
The nanostructured ZrO2 coating shows the highest hardnessvalues but also the highest specific wear rate, one to two ordersof magnitude higher than the other prepared coatings, inparticular those deposited with MA powders.
�
Adhesive wear mechanism was observed in the worn surface ofthe nickel coating and coatings prepared by Dual deposition,
F. Fernandes et al. / Wear 303 (2013) 591–601 601
while abrasive wear was observed in ZrO2 coating and coatingsdeposited using MA powders.
Acknowledgments
The authors wish to express their sincere thanks to thePortuguese Foundation for the Science and Technology (FCT),through COMPETE program from QREN and to FEDER, for financialsupport in the aim of the Project number “13545”, as well as for theGrant (SFRH/BD/68740/2010).
References
[1] S. Harsha, D. Dwivedi, A. Agarwal, Influence of CrC addition in Ni–Cr–Si–Bflame sprayed coatings on microstructure, microhardness and wear behaviour,International Journal of Advanced Manufacturing Technology 38 (2008)93–101.
[2] J.M. Miguel, J.M. Guilemany, S. Vizcaino, Tribological study of NiCrBSi coatingobtained by different processes, Tribology International 36 (2003) 181–187.
[3] J.C. Miranda, A. Ramalho, Abrasion resistance of thermal sprayed compositecoatings with a nickel alloy matrix and a WC hard phase. Effect of depositiontechnique and re-melting, Tribology Letters 11 (2001) 37–48.
[4] A.K. Basak, S. Achanta, J.P. Celis, M. Vardavoulias, P. Matteazzi, Structure andmechanical properties of plasma sprayed nanostructured alumina andFeCuAl–alumina cermet coatings, Surface and Coatings Technology 202(2008) 2368–2373.
[5] Y. Wu, S. Hong, J. Zhang, Z. He, W. Guo, Q. Wang, G. Li, Microstructure andcavitation erosion behavior of WC–Co–Cr coating on 1Cr18Ni9Ti stainless steelby HVOF thermal spraying, International Journal of Refractory Metals andHard Materials 32 (2012) 21–26.
[6] R.A. Mahesh, R. Jayaganthan, S. Prakash, Oxidation behavior of HVOF sprayedNi–5Al coatings deposited on Ni- and Fe-based superalloys under cycliccondition, Materials Science and Engineering: A 475 (2008) 327–335.
[7] H. Liao, B. Normand, C. Coddet, Influence of coating microstructure on theabrasive wear resistance of WC/Co cermet coatings, Surface and CoatingsTechnology 124 (2000) 235–242.
[8] C.-J. Li, C.-X. Li, Y.-Z. Xing, M. Gao, G.-J. Yang, Influence of YSZ electrolytethickness on the characteristics of plasma-sprayed cermet supported tubularSOFC, Solid State Ionics 177 (2006) 2065–2069.
[9] M.H. Staia, T. Valente, C. Bartuli, D.B. Lewis, C.P. Constable, A. Roman, J. Lesage,D. Chicot, G. Mesmacque, Part II: tribological performance of Cr3C2–25% NiCrreactive plasma sprayed coatings deposited at different pressures, Surface andCoatings Technology 146–147 (2001) 563–570.
[10] G.J. Gibbons, R.G. Hansell, Thermal-sprayed coatings on aluminium for mouldtool protection and upgrade, Journal of Materials Processing Technology 204(2008) 184–191.
[11] J.H. Chang, C.P. Chang, J.M. Chou, R.I. Hsieh, J.L. Lee, Microstructure and bondingbehavior on the interface of an induction-melted Ni-based alloy coating and AISI4140 steel substrate, Surface and Coatings Technology 204 (2010) 3173–3181.
[12] W. Li, Y. Li, C. Sun, Z. Hu, T. Liang, W. Lai, Microstructural characteristics anddegradation mechanism of the NiCrAlY/CrN/DSM11 system during thermalexposure at 1100 1C, Journal of Alloys and Compounds 506 (2010) 77–84.
[13] C. Guo, J. Zhou, J. Chen, J. Zhao, Y. Yu, H. Zhou, High temperature wearresistance of laser cladding NiCrBSi and NiCrBSi/WC–Ni composite coatings,Wear 270 (2011) 492–498.
[14] F. Fernandes, A. Cavaleiro, A. Loureiro, Oxidation behavior of Ni-based coatingsdeposited by PTA on gray cast iron, Surface and Coatings Technology 207(2012) 196–203.
[15] Q.Y. Hou, Z. Huang, J.T. Wang, Influence of nano-Al2O3 particles on the micro-structure and wear resistance of the nickel-based alloy coating deposited byplasma transferred arc overlay welding, Surface and Coatings Technology 205(2011) 2806–2812.
[16] H.-y. Wang, D.-w. Zuo, M.-d. Wang, G.-f. Sun, H. Miao, Y.-l. Sun, High temperaturefrictional wear behaviors of nano-particle reinforced NiCoCrAlY cladded coatings,Transactions of Nonferrous Metals Society of China 21 (2011) 1322–1328.
[17] W. Wang, H.T. Guo, J.P. Gao, X.H. Dong, Q.X. Qin, XPS, UPS and ESR studies onthe interfacial interaction in Ni–ZrO2 composite plating, Journal of MaterialsScience 35 (2000) 1495–1499.
[18] K.F. Zhang, S. Ding, G.F. Wang, Different superplastic deformation behavior ofnanocrystalline Ni and ZrO2/Ni nanocomposite, Materials Letters 62 (2008)719–722.
[19] F.Y. Hou, W. Wang, H.T. Guo, Effect of the dispersibility of ZrO2 nanoparticles inNi–ZrO2 electroplated nanocomposite coatings on the mechanical propertiesof nanocomposite coatings, Applied Surface Science 252 (2006) 3812–3817.
[20] W. Wang, F.Y. Hou, H. Wang, H.T. Guo, Fabrication and characterizationof Ni–ZrO2 composite nano-coatings by pulse electrodeposition, ScriptaMaterialia 53 (2005) 613–618.
[21] A. Ramalho, P.V. Antunes, Reciprocating wear test of dental composites againsthuman teeth and glass, Wear 263 (2007) 1095–1104.
[22] A. Ramalho, A reliability model for friction and wear experimental data, Wear269 (2010) 213–223.
[23] J.G. La Barbera-Sosa, Y.Y. Santana, E. Moreno, N. Cuadrado, J. Caro, P.O. Renault,E. Le Bourhis, M.H. Staia, E.S. Puchi-Cabrera, Effect of spraying distance on themicrostructure and mechanical properties of a Colmonoy 88 alloy depositedby HVOF thermal spraying, Surface and Coatings Technology 205 (2010)1799–1806.
[24] X.B. Zhao and Z.H. Ye, Microstructure and wear resistance of molybdenumbased amorphous nanocrystalline alloy coating fabricated by atmosphericplasma spraying, Surface and Coatings Technology http://dx.doi.org/10.1016/j.surfcoat.2012.05.127, in press.
[25] T. Sidhu, S. Prakash, R.D. Agrawal, Studies on the properties of high-velocityoxy-fuel thermal spray coatings for gigher temperature applications, MaterialsScience 41 (2005) 805–823.
[26] J.C. Fang, H.P. Zeng, W.J. Xu, Z.Y. Zhao, L. Wang, Prediction of in-flight particlebehaviors in plasma spraying, World Academy of Materials and ManufacturingEngineering 18 (2006) 283–286.
[27] O. Sarikaya, Effect of some parameters on microstructure and hardness ofalumina coatings prepared by the air plasma spraying process, Surface andCoatings Technology 190 (2005) 388–393.
[28] P. Fauchais, Understanding plasma spraying, Journal of Physics D: AppliedPhysics 37 (2004) R86–R108.
[29] H. Chen, S. Lee, X. Zheng, C. Ding, Evaluation of unlubricated wear propertiesof plasma-sprayed nanostructured and conventional zirconia coatings by SRVtester, Wear 260 (2006) 1053–1060.
[30] L.M. Chang, M.Z. An, H.F. Guo, S.Y. Shi, Microstructure and properties of Ni–Co/nano-Al2O3 composite coatings by pulse reversal current electrodeposition,Applied Surface Science 253 (2006) 2132–2137.
F. Fernandes, A. Loureiro, T. Polcar, A. Cavaleiro, "The effect of increasing V
content on the structure, mechanical properties and oxidation resistance of Ti-Si-V-
N films deposited by DC reactive magnetron sputtering, Applied Surface Science,
289, (2014) 114-123.
TpD
Fa
b
c
a
ARRAA
KTSMOV
1
ucadiamdvolpb[
f
0h
Applied Surface Science 289 (2014) 114– 123
Contents lists available at ScienceDirect
Applied Surface Science
jou rn al h omepa g e: www.elsev ier .com/ locate /apsusc
he effect of increasing V content on the structure, mechanicalroperties and oxidation resistance of Ti–Si–V–N films deposited byC reactive magnetron sputtering
. Fernandesa,∗, A. Loureiroa, T. Polcarb,c, A. Cavaleiroa
CEMUC—Department of Mechanical Engineering, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, PortugalDepartment of Control Engineering, Czech Technical University in Prague, Technicka 2, Prague 6 166 27, Czech RepublicnCATS, University of Southampton, Highfield Campus, SO17 1BJ Southampton, UK
r t i c l e i n f o
rticle history:eceived 26 June 2013eceived in revised form 18 October 2013ccepted 19 October 2013vailable online 28 October 2013
eywords:iSi(V)N filmstructureechanical properties
a b s t r a c t
In the last years, vanadium rich films have been introduced as possible candidates for self-lubricationat high temperatures, based on the formation of V2O5 oxide. The aim of this investigation was to studythe effect of V additions on the structure, mechanical properties and oxidation resistance of Ti–Si–V–Ncoatings deposited by DC reactive magnetron sputtering. The results achieved for TiSiVN films were com-pared and discussed in relation to TiN and TiSiN films prepared as reference. All coatings presented a fccNaCl-type structure. A shift of the diffraction peaks to higher angles with increasing Si and V contentssuggested the formation of a substitutional solid solution in TiN phase. Hardness and Young’s modulusof the coatings were similar regardless on V content. The onset of oxidation of the films decreased sig-nificantly to 500 ◦C when V was added into the films; this behaviour was independent of the Si and V
xidation resistanceanadium oxide
contents. The thermogravimetric isothermal curves of TiSiVN coatings oxidized at temperatures belowthe melting point of �-V2O5 (∼685 ◦C) showed two stages: at an early stage, the weight increase overtime is linear, whilst, in the second stage, a parabolic evolution can be fitted to the experimental data. Athigher temperatures only a parabolic evolution was fitted. �-V2O5 was the main phase detected at theoxidized surface of the coatings. Reduction of �-V2O5 to �-V2O5 phase occurred for temperatures above
its melting point.
. Introduction
High-speed cutting and dry machining processes without these of environmental harmful lubrication requires milling toolsapable to withstand severe conditions of wear, friction, oxidationnd corrosion [1,2]. Solid lubricants coatings, such as WC/C, MoS2,iamond-like carbon (DLC), hex-BN as well as their combinations
n nanocrystaline or multilayer structures, have been successfullypplied in such parts in order to increase their lifetime and perfor-ance for various tribological applications. However, considerable
egradation of the tribological effectiveness of these coatings at ele-ated temperature has been reported due to their low resistance toxidation [3–5]. To overcome this shortcoming, a new concept ofubrication based on the formation of lubricious oxides has been
roposed. This is the case of the so-called Magnéli oxide phasesased on Ti, Si, Mo, W and V, which have easy shear able planes3,6]. Among these elements, particular attention has been given
to the vanadium-containing coatings (Magnéli phases VnO3n−1),which showed interesting tribological properties in the temper-ature range 500–700 ◦C [3,7–12]. Dissimilar series of V-based hardcoatings have been developed, such as ternary CrVN [13], (V,Ti)N[14], multilayer AlN/VN [15] and quaternary single layer or mul-tilayered CrAlVN [16,17] and TiAlVN [2,8,18,19]. However, sinceother ternary coatings systems, based on (TiX)N with X = B, Cr, Al,Si, Cr, etc, have been attractive for advanced hard coating materi-als, the effect of vanadium doping to these systems should also beconsidered.
Among those ternary systems, most of the research works havebeen focused on TiSiN coatings deposited by CVD and/or PVD tech-niques [20,21]. Depending on the deposition conditions, these coat-ings could be deposited with a nanocomposite structure, consistingof nano-sized TiN crystallites surrounded by an amorphous matrixof Si3N4 displaying very high hardness values [22], compared to anyother ternary (TiX)N compound, with extremely good oxidationresistance [23,24]. In other deposition conditions, substitutional
solid solution of Si in TiN structure (not predicted by the Ti–Si–Nphase diagram) could be achieved [22,25,26]. Relatively to thefriction coefficient of Ti–Si–N coatings, the studies have reportedvalues in the range [0.6–1.3] at room and high temperatures
F. Fernandes et al. / Applied Surface Science 289 (2014) 114– 123 115
Table 1Sample designation and deposition parameters of the coatings.
Sample Target power density (W/cm2) Deposition time (min) Pellets of V at the Ti target
Ti TiSi2
Interlayer Ti 10 – 5 –Interlayer TiN – 15 –
TiSiN S1-0 6.5 1 210 0S1-4 4
TiSiVN S1-8 8S1-12 12
TiSiN S2-0 6 1.5 193 0S2-4 4
TiSiVN S2-8 8S2-12 12
TiSiN S3-0 5 2.5 153 0S3-4 4
TiSiVN S3-8 8
[cooe
vmiwsa
2
situ1ecctotwswdtStTr(u1
c(las
S3-12
TiN 10 –
27–29]. Since the addition of V successfully decreased the frictionoefficient of binary and ternary systems (down to reported valuesf 0.2–0.3 at temperatures between 550 and 700 ◦C), similar studiesn the effect of V-addition should be performed on TiSiN coatingsxhibiting unique mechanical properties and oxidation resistance.
This paper reports the first results on the effect of increasinganadium content to Ti–Si–V–N films deposited by DC reactiveagnetron sputtering. It is focused on coating structure, mechan-
cal properties and oxidation resistance. Comparison of the resultsith those achieved for reference TiN and TiSiN coatings is pre-
ented. The results of this study will support further research aimedt the tribological behaviour of the coatings at high temperatures.
. Experimental procedure
Three series of TiSiN films with different Si contents, eacheries with increasing V contents (TiSiVN coatings), were depositedn a d.c. reactive magnetron sputtering machine equipped withwo rectangular (100 × 200 mm) magnetron cathodes working innbalanced mode. A high purity Ti (99.9%) target, with 18 holes of0 mm in diameter (uniformly distributed throughout the prefer-ntial erosion zone of the target), and a high purity TiSi2 (99.9%)omposite target were used in the depositions. The different sili-on contents were achieved by changing the power density appliedo each target. The V content was varied by changing the numberf high purity rods of Ti and V (with 10 mm in diameter) placed inhe holes of the Ti target. 4, 8 and 12 cylindrical pieces of vanadiumere used. In all the cases the total power applied to the targets was
et to 1500 W. To serve as reference, a stoichiometric TiN coatingas deposited from the Ti target. Hereinafter the coatings will beesignated as Sx-y, where x is related to the specific power appliedo the TiSi2 target (see Table 1), and therefore associated to thei content on the coatings, and y the number of V rods used inhe depositions, giving an indication of V content in each series ofiSiVN films. Thus, denomination S2-0, S2-4, S2-8 and S2-12 rep-esents coatings from the same series, i.e. with identical Si contentTiSi2 target power 1.5 W/cm2), with increasing V content from 0p to a maximum content achieved for the coating produced using2 vanadium pellets embedded into the Ti target.
Polished high-speed steel (AISI M2) (Ø 20 × 3 mm, for mechani-al properties measurements), FeCrAl alloy and alumina substrates
10 × 10 × 1 mm, for oxidation tests and structural analysis), stain-ess steel discs (Ø 20 × 1 mm, for residual stress measurements)nd (1 1 1) silicon samples (10 × 10 × 0.8 mm, for thickness mea-urements and chemical composition evaluation) were used as
1273 –
substrates. Prior to the depositions, all the substrates were ultra-sonically cleaned in acetone for 15 min and alcohol for 10 min. Thesubstrates were mounted in a substrate holder (which revolvedwith 18 rev/min around the centre axis) giving a target to sub-strate distance of 175 mm. Prior to deposition, the chamber wasevacuated down to 8.7 × 10−4 Pa and the substrates were etchedwith Ar ion sputtering during 1 h with a bias voltage of −650 V toremove any surface contaminants. In order to enhance the adhesionof the coatings, Ti and TiN adhesion layers of approximately 0.24and 0.45 �m, respectively, were deposited on the substrates beforeTiSi(V)N coatings. In all the depositions, the total working gas pres-sure was kept constant at 0.3 Pa, using approximately 30 sccm ofAr and 17 sccm of N2. The depositions were performed with a neg-ative substrate bias of 50 V. The deposition time was set in order toobtain films with approximately 2.5 �m of total thickness (includ-ing interlayers). The deposition parameters are shown in Table 1.
The chemical composition of the coatings was evaluated byelectron probe microanalysis (EPMA—Cameca SX 50). Crystallo-graphic structure was investigated by X-ray diffraction (X’ PertPro MPD diffractometer) using a grazing incidence angle of 1◦ andCu K�1 radiation (� = 1.54060 A). The XRD spectra were fitted byusing a pseudo-Voigt function to calculate either the full widthat half maximum (FWHM) and the peak position (2�). The frac-ture cross-section morphology and the thickness of the films wereinvestigated by scanning electron microscopy (SEM). Further, pro-filometer was used to confirm coating thickness and to evaluatesurface roughness.
The hardness and Young’s modulus of films were measured ina nano-indentation equipment (Micro Materials NanoTest) using aBerkovich diamond pyramid indenter. In order to avoid the effectof the substrate, the applied load (10 mN) was selected to keepthe indentation depth less than 10% of the coating’ thickness. 32measurements were performed in each sample. The level of resid-ual stresses was calculated through the Stoney equation [30] bymeasuring the bulk deflection of the film-substrate body.
The oxidation resistance of the coatings was evaluated by ther-mogravimetric analysis (TGA) using industrial air (99.99% purity).The films deposited on alumina substrates were firstly heated witha constant temperature ramp of 20 ◦C/min from room temperatureup to 1200 ◦C, in order to determine the onset point of oxidation.Then, coated specimens of FeCrAl alloy were subjected to isother-
mal tests at different selected temperatures and time. The weightgain of the samples was evaluated at regular 2 s intervals using amicrobalance with an accuracy of 0.01 mg. The air flux used was50 ml/min and the heating rate up to the isothermal temperature
116 F. Fernandes et al. / Applied Surface
F(
wtwst
3
3
3
bnpcro
ctSadcrtco
ig. 1. XRD patterns of: (a) TiSiN coatings with Si content ranging from 0 to 12.6 at.%,b) coating S1-0 with V additions.
as 20 ◦C/min. After oxidation tests, the surface morphology ofhe specimens was examined by scanning electron microscopyith energy dispersive X-ray spectroscopy (SEM–EDS) and their
tructure was evaluated by XRD diffraction. The oxide products onhe surface were characterized by Raman spectroscopy.
. Results and discussion
.1. Coatings characterization
.1.1. Chemical composition and microstructureThe elemental chemical composition of the coatings determined
y EPMA is shown in Table 2. As could be expected, the simulta-eous decrease in the power density of Ti target and increase in theower density of TiSi2 target gave rise to higher Si content in theoatings. Only insignificant variations were detected on the Si/Tiatios with incorporation of V in the Ti target. The nitrogen contentf the coatings was close to 50 at.% regardless of Si and V.
The corresponding X-ray diffraction patterns of the as-depositedoatings are displayed in Fig. 1. The influence of increasing V con-ent on TiSiN coatings is shown in Fig. 1b) for the film with loweri content, as representative of all series containing vanadium. Inll graphs TiN film pattern is shown as a reference. In all cases, theiffraction peaks could be generally assigned to the fcc NaCl-typerystalline TiN phase. The XRD pattern of the TiN reference coating
eveals that all peaks are shifted to lower diffraction angles in rela-ion to the standard ICDD card of TiN, probably indicating residualompressive stresses [31]. For TiSiN coatings with 3.8 and 6.7 at.%f Si, (2 0 0) peak vanishes and the sharpening of the peaks suggests
Science 289 (2014) 114– 123
larger grain size. Based on Scherrer’ equation, grain sizes of 16 and24 nm were calculated for TiN and coatings with low Si content,respectively. With further increase of the Si content (12.6 at.% ofSi) a strong decrease of the peaks intensity, as well as their signifi-cant broadening (a grain size of 6 nm was calculated), are observedsuggesting a decrease of crystallinity. With increasing Si content,the fcc peaks are shifted to higher angles, behaviour associated toa smaller unit cell. This can be explained by the presence of Si insolid solution, since its smaller atomic radius promotes the con-traction of the TiN lattice. Therefore, in this case the stoichiometricphases (TiN + Si3N4) do not segregate and nanocomposite structureis not formed. Such behaviour could be explained by the low depo-sition temperature together with low substrate ion bombardment,which do not provide the necessary mobility of the species for thenanocomposite formation [22]. This observation is in agreementwith the results from the literature, where similar solid solutionsand changes in the film orientation have been observed with siliconincorporation [21,25,32].
The coatings with V displayed similar XRD patterns and peaksto TiSiN. No significant changes in the peaks intensities or widthswere observed, except for the highest Si content films where asmall improvement in the crystallinity was detected. All the vana-dium containing films had the XRD peaks shifted to higher angles,behaviour, once again, related to a smaller unit cell due to the sub-stitution of Ti by the smaller V atoms. Similar trend was observedby Pfeiler et al. [2] for the TiAlVN system.
Fig. 2 displays typical surface and fracture cross section-micrographs of TiSiVN coatings. All coatings showed a typicalcolumnar structure. As Si content was increased (see Fig. 2b)) forcoating S1-0), the columns size and the surface roughness werefirstly reduced, for silicon contents of 3.8 and 6.7 at.%, and thenincreased for the highest silicon content, which is an oppositetrend than that observed for the grain size. The addition of V didnot change the global type of columnar morphology and the coat-ing cross-section was similar. The only exception was the strongincrease of the size of the columns as well as the surface roughness(see Fig. 2c) for coating S1-12) for low silicon content coatings.
3.1.2. Hardness, Young’s modulus and residual stressesThe dependence of the hardness and Young’ modulus on the
Si and V content is shown in Fig. 3. As the silicon content in theTiSiN films increases, the hardness of coatings reached a maximumvalue of 27 GPa at a Si content of 6.7 at%; then it dropped with fur-ther increasing Si content to a lower value than the reference TiN.Although the grain grow is observed with increasing Si additionsup to 6.7 at% of Si, the hardening of coatings is mostly result of thenitride lattice distortion, which improves the resistance to plasticdeformation (solid solution hardening). Residual stresses shouldnot be a factor of hardness differentiation between the coatings,since all films have approximately the same level of compressiveresidual stresses, with any Si addition (∼3 GPa). This result also sup-ports what was stated above that the shift of the peaks to the rightwith Si additions is exclusively due to substitutional solid solu-tion. The decrease in hardness with further Si addition is due tothe loss of crystallinity associated with the lower degree of phasenitriding regardless on the lower grain size [26]. Young’s modu-lus progressively decreases with increasing Si content, which isrelated to changes in the binding energy between ions due to Siincorporations. Similar hardness and Young’s modulus evolutionas a function of Si content has been reported in Refs [25,33].
Concerning the effect of V addition on the hardness and Young’modulus, similar evolution was observed for all films (see e.g.
Fig. 3b) for S1-y films). Although the variations are very small andoften within the error bars of the hardness values, general trend isa slight increase in hardness followed by drop for the highest vana-dium content. Coatings with the higher silicon content and with
F. Fernandes et al. / Applied Surface Science 289 (2014) 114– 123 117
Table 2Chemical composition of the dissimilar coatings in at.%.
anadium additions displayed lower hardness and Young’s modu-us values as compared to the other V rich coatings. The hardnessnhance with V additions is probably due to the presence of V inolid solution. In fact, the similar level of residual stresses as func-ion of V content measured, which revealed to be from the rangef 3–4 GPa, and the observed shift of peaks to higher angles with Vncorporation, supports the previous affirmation.
.2. Continuous and isothermal oxidation in air
The effect of increasing Si and V content on the onset point ofxidation of the TiN and TiSiN systems, respectively, are shown inig. 4. Silicon incorporation strongly improved the onset of oxida-ion of the coatings, with the highest silicon content (S3-0) showinghe highest oxidation resistance. Kacsich et al. [34,35] reported thathe high oxidation resistance of these coatings was due to the pres-nce of a protective silicon-rich oxide layer. This layer acted as anfficient diffusion barrier against oxygen and metal ions diffusion;hus, it protected the coating from further oxidation. Our resultshowed that the onset point of oxidation of films decreased sig-ificantly down to a temperature of approximately 500 ◦C (lowerhan TiN) for all coatings containing vanadium. Interestingly, thisehaviour was independent of Si content of the films and, there-ore, the vanadium incorporation interfered with the diffusion ofilicon and the consequent formation of a continuous protectiveilicon oxide layer. Furthermore, all V-containing coatings wereompletely oxidized at a lower temperature than TiN film. Lewist al. [7], Zhou et al. [36], and Franz et al. [37], who studied the effectf V doping on the oxidation behaviour of multilayered TiAlN/VNnd CrAlVN coatings, respectively, also observed a similar decreasen the onset point of oxidation. In order to better characterize thexidation behaviour of the V rich coatings, isothermal oxidationests were conducted at selected temperatures using S2-8 coatingeposited onto FeCrAl alloy substrates. The isothermal curves wereompared with those of TiSiN coating with similar Si content andith TiN (see Fig. 5).
Comparing the mass gain of the coatings and taking into accounthe testing temperature in each case, it is possible to conclude that-free TiSiN coating (S2-0 tested at 900 ◦C) is much more resistant
o oxidation. This coating exhibits a typical parabolic oxidationeight gain as a function of time indicating the formation ofrotective silicon-rich oxide layer, as referred to above. Similarype of evolution is exhibited by TiN coating; however, in this case
he titanium oxide scale does not protect the coatings and, thus,t is effective only at low isothermal temperature [38]. It shoulde remarked that during continuous oxidation test up to 1200 ◦C,t 900 ◦C TiN coatings was already half consumed. Regarding the
isothermal curves of TiSiN coating with V additions (S2-8), it canbe concluded that, independently of the isothermal temperatureand time of exposure, their mass gain is always much higher thanTiN and V-free TiSiN coatings (S2-0). The isothermal curves testedat 550 and 600 ◦C showed two steps: at an early stage, the weightgain is rapidly increasing almost linearly up to approximately0.1 mg/cm2 (particularly at 600 ◦C), whereas in the second stage,a parabolic evolution can be fitted. Isothermal annealing of S2-8at 700 ◦C shows significant increase of mass gain following theparabolic evolution.
3.2.1. Surface structureAs it would be expected, both TiN annealed at 600 ◦C and S2-0
coating annealed at 900 ◦C exhibited several intense peaks belong-ing to the (1 1 0), (1 0 1), (2 0 0), (1 1 1) and (2 1 0) of a tetragonalphase related to rutile/TiO2 (ICDD card 76-0649) (not shown here).Weak peaks assigned to TiN (ICDD card 87-0628) were observedin the XRD spectra of the S2-0 coating, which was in a good agree-ment with its annealing curve, since only partial oxidation of thefilm was expected. Moreover, diffraction peaks associated with Sioxide were not detected on the oxidized surface of S2-0 film sug-gesting an amorphous character of the protective oxide layer. Thisresult is in agreement with literature, where similar observationwas reported [34,35].
XRD patterns of S2-8 oxidized coatings in isothermal tests areshown in Fig. 6. The main phases detected by XRD were Ti–V–Oand V–O oxides; however, depending on the isothermal temper-ature and time, different orientations and peaks assigned to V–Owere perceived. After annealing at 550 ◦C during 1 h, the follow-ing phases could be identified: Ti(V)O2 (ICDD card 77-0332) and�-V2O5 (ICDD card 41-1426) with a preferred orientation follow-ing the (0 0 1) plane (peak at 20.3◦), together with the f.c.c. nitridefrom the coating. The same phases were detected at 600 ◦C during10 min, however, �-V2O5 phase showed random orientation. Fur-ther increase in isothermal time (annealing at 600 ◦C during 30 min)resulted in loss of TiN peaks and the orientation of the �-V2O5 phaseis again following (0 0 1) plane. The absence of TiN peaks is in goodagreement with the isothermal curve shown in Fig. 5. Progressiveoxidation is demonstrated in the intensity of (0 0 1) reflection after30 min of annealing, where the amount of �-V2O5 is clearly higher.These changes in phase proportion and orientation have alreadybeen observed by several authors during the oxidation of TiAlN/VNmultilayer and TiAlVN single layered coatings [1,7,8,36]. Increas-
ing the heat treatment to 700 ◦C, which is temperature higher thanthe melting point of �-V2O5 (∼685 ◦C [10,36]), resulted in the ran-dom orientation of this phase; moreover, a sharp peak at 2� = 12.8◦
identified as �-V2O5 (ICDD card 45-1074) with (0 0 2) preferential
118 F. Fernandes et al. / Applied Surface Science 289 (2014) 114– 123
Fig. 3. (a) Effect of Si content on hardness and Young’s modulus for the TiN system.(b) Effect of V content on hardness and Young’s modulus for coating S1-0.
orientation appearing [39]. �-V2O5 phase is a result of loss of Ofrom �-V2O5, due to a reduction process [40]. We should empha-size here that vanadium oxides observed on the oxidized surface areoptimal from the tribological point of view. Vanadium oxide couldacts as low-friction solid lubricant or, for higher temperatures, asliquid lubricant. Therefore the drop in onset oxidation tempera-ture should not be considered as a significant drawback of TiSiVNcoatings.
3.2.2. Surface morphology and characterizationTypical SEM surface micrographs of the S2-0 and V-rich coating
(S2-8) isothermally oxidized at different temperatures and times
Fig. 4. Thermal gravimetric oxidation rate of coatings deposited on Al2O3 substratesusing the linear-temperature ramp (RT to 1200 ◦C at 20 ◦C/min).
F. Fernandes et al. / Applied Surface
Fig. 5. Thermo gravimetric isothermal analysis of coatings exposed at differenttemperatures.
Ft
awtft
Fa
ig. 6. X-ray diffraction patterns of oxidized surface of: coating S2-8 annealed atemperatures ranging from 550 to 700 ◦C.
re shown in Figs. 7 and 8, respectively. EDS and Raman analysesere carried out on the oxidized surface of the coatings to identify
heir composition. On the oxidized surface of the S2-0 film, two dif-erent zones could be detected, a dark grey phase evenly distributedhroughout the surface and white islands. EDS analysis showed that
ig. 7. SEM observation of surface morphology of coating S2-0, after 1 h oxidationt 900 ◦C.
Science 289 (2014) 114– 123 119
the white oxide was rich in Ti (crystals of TiO2), and the dark greyphase composed mainly by Si and Ti. The Raman spectra of thesezones are plotted in Fig. 9. As can be observed the light grey phasedisplays Raman active modes at 152, 253, 454 and 610 cm−1 [3,23]assigned to rutile, confirming its presence in the oxidized coatingdetected by XRD. The same phase, with much less intensity, wasdetected for the dark grey phase; however, strong additional peaksassigned to anatase (TiO2) were observed. This finding corroboratesthe report of Pilloud et al. [23], who showed that anatase Ramanbands increased with the silicon content in TiSiN films, whereasthat of rutile decreased. As only strong peaks of rutile were detectedby XRD, this indicates that the amount of anatase should be smalland, therefore, not detectable by XRD diffraction. Since Raman pen-etration depth is relatively limited, the absence of silicon oxide inRaman spectra indicated that Si–O was below theTiO2 phase iden-tified by XRD. This finding corroborates the results of Kacsich et al.[34,35]; they observed titanium oxide layer formed at outmost sur-face and silicon oxide sublayer (TiSiN coatings).
The analysis of the oxidized surface of S2-8 coating reveals thatthe surface oxide morphology is different from that of coating S2-0. It displays a floret-like structure formed by light and dark greyzones. At 550 ◦C these phases are tiny distributed throughout thesurface, being difficult to clearly identify the boundaries betweenthem. At a temperature of 600 ◦C, the separation between bothzones is evident with the shape of darker zone suggesting some typeof dendritic growth. With the increase of the isothermal time, anincrease in the area covered by the grey dark phase was observed.The temperature 700 ◦C led to a different oxide morphology, i.e.the appearance of a black zone in some rosettes and the segrega-tion of a white phase to the boundaries of the floret-like structure.This change in the microstructure can be associated to the meltingof �-V2O5, which originates a smoother surface (Fig. 8d)). Fig. 10shows examples of EDS spectra of points identified in Fig. 8 (phases1 and 2). As can be observed, Ti K� peak (4.931 keV) overlaps the VK� peak (4.952 keV) and, therefore, V K� peak should be taken inconsideration for analyzing V importance. In order to identify thedifferent oxide phases marked in Fig. 8, the ratio between the peaksintensities of V K� (Si K�) and Ti K� from EDS analyses are sum-marized in Table 3. To a relative increase in this ratio correspondsthe preferential formation of the oxide of that element. EDS analy-ses of points 1 and 2 reveal similar compositions for the grey lightand dark zones, suggesting the presence of oxides containing Ti, Siand V. However, darker phase has clearly a higher V content (seeTable 3). The Raman spectrum taken at grey dark phase (see Fig. 11)shows the presence of intense peaks assigned to �-V2O5 [3,8,10]and small peaks related to rutile (TiO2). On light grey phase thespectrum is much less defined, with broader bands. �-V2O5 peaksalmost vanished whereas rutile bands are enhanced. Low crystal-lized Ti–O phases give very similar Raman spectra, which evolve,after annealing, either as anatase or rutile [41,42] phases. In sum-mary, the dark grey phase can be assigned mainly to �-V2O5, andsmall quantities of TiO2, probably as a bilayer with �-V2O5 on thetop and TiO2 underneath, as suggested by EDS measurements. Infact, the high ratio of V/Ti of point 3 marked in Fig. 8, allows iden-tifying a V-rich oxide, on the surrounding regions of the dark greyphase. The light grey phase should be attributed to low crystallizedTi(V)O2. These results agree to the XRD results acquired on the oxi-dized surface of the films where both Ti(V)O2 and �-V2O5 phaseswere indexed. When the isothermal time was prolonged from 10 to30 min at 600 ◦C, the increase of the intensity of the �-V2O5 peakobserved in the XRD patterns was is in a good agreement with ahigher amount of grey dark phase observed on the oxidized surface
morphology.
Similarly to S2-0 coating signals from Si–O oxide were neitherdetected by XRD nor by Raman spectroscopy. However, accordingto EDS analysis this oxide should be present. As these signals are
120 F. Fernandes et al. / Applied Surface Science 289 (2014) 114– 123
Fig. 8. Typical surface morphology of oxidized coatings: (a) coating S2-8 exposed to 55exposed to 600 ◦C during 30 min, (d) coating S2-8 exposed to 700 ◦C during 10 min.
Fig. 9. Raman spectra of dark and light grey phases identified in Fig. 7.
TR
able 3atio between the peaks intensities of V K� (Si K�) and Ti K� from EDS analyses of points
1 2
RatiosV K�/Ti K� 0.04 0.1
Si K�/Ti K� 0.77 0.99
0 ◦C during 1 h, (b) coating S2-8 exposed to 600 ◦C during 10 min, (c) coating S2-8
coming from a sub-surface layer, an oxide layer rich in Ti and Sishould coexist with a global chemical composition similar to thatof light grey phase.
From the three different morphological zones of sampleannealed at 700 ◦C for 10 min (Fig. 8d)), EDS analysis showed thatthe two grey phases were mainly rich in vanadium and oxygen. Sig-nals of elements from the substrate were detected in the dark phase,particularly Al, which in conjunction with XRD results, allowedidentifying it as �-V2O5. In fact, it is well known, that the loss of Ofrom �-V2O5, due to a reduction process, leads to its transformationin �-V2O5 [40,43]. The presence of highly oxygen reactive elements,such as aluminium, in the substrate can reduce the molten �-V2O5,subtracting O and promoting the formation of �-V2O5. The Ramanspectra of the dark and light grey phases plotted in Fig. 12 showsimilar Raman patterns with peaks related to V2O5. Both �- and�-V2O5 display similar Raman spectra [44]. The segregated whiteareas located at the boundaries of the grey phase (see ratio Si/Ti ofpoint 5) were rich in Si and Ti, suggesting to be silicon and titanium
oxides.
A good correlation was found between the isothermal oxida-tion curve, XRD measurements and surface morphology of S2-8coating. At temperatures below the melting point of �-V2O5, the
marked in Fig. 8.
Points
3 4 5 6
0.15 0.63 0.01 0.680.19 0.58 0.49 1.00
F. Fernandes et al. / Applied Surface Science 289 (2014) 114– 123 121
F
igit�igfslt�aosop
ig. 10. Energy dispersive X-ray analysis (EDAX) of points 1 and 2 identified in Fig. 8.
sothermal oxidation curve starts with a linear increase in the massain and then changes to a parabolic law. The XRD spectra of spec-men S2-8 oxidized during 10 min at 600 ◦C, in the first stage ofhe oxidation process, revealed the presence of intense peaks of-V2O5. Further increase in isothermal time to 30 min showed an
ncreased amount of �-V2O5, as confirmed by the increase of therey dark phase, above attributed to V-rich zones. XRD patternor this sample also showed an increase in Ti(V)O2 peaks inten-ity suggesting a thickening of the oxide scale. Therefore, the firstinear part of the isothermal curve is due to the oxidation of bothitanium and vanadium to form Ti(V)O2 and �-V2O5, respectively.-V2O5 is a non-protective oxide and Ti(V)O2 is still not in enoughmounts to form a continuous protective scale, reason why a linear
xidation rate is observed. For longer oxidation times, the progres-ive thickening of the Ti-rich oxide scale will make difficult thexygen diffusion inward retarding the reaction and promoting thearabolic behaviour. However, the protection is not so efficient
Fig. 11. Raman spectra of phase 1 and 2 identified in Fig. 8.
Fig. 12. Raman spectra of phase 4 and 5 identified in Fig. 8.
as in TiN and/or Ti–Si–N samples (see Fig. 5) due to the disrup-tion induced by the outward/lateral diffusion of V ions through theoxide scale. In any condition the protective Si–O layer could be alsoformed. Similar isothermal evolution was reported by Zhou et al.[36], which during isothermal oxidation of multilayered TiAlN/VNcoatings have detected �-V2O5 at the initial stage of oxidation.In their system, the mass gain curves also showed initially a lin-ear evolution, followed by a parabolic growth, similar to whatwas observed here for TiSiVN films. Keller and Douglass [45] alsoshowed that pure vanadium oxidizes linearly, whereas the additionof different elements (Al, Cr, Ti) could transform the initial linearmass gain for a parabolic one due to the modification of the formedoxide layer.
At 700 ◦C, temperature higher than the melting point of �-V2O5, very high oxidation rates were measured and high amountsof V-oxides were detected by XRD. The extremely high parabolicevolution suggests a very high diffusion rate of the reactive ionsthrough the oxide scales.
A comparison between TiSiVN with CrAlVN and TiAlVN (insingle layered or multilayered configuration) coatings, reveals alower onset point of oxidation of approximately 100 ◦C. In CrAlVNand TiAlVN coatings the onset point of oxidation was close to∼600 ◦C [8,36,37]. In the case of the TiAlVN films, signals of lubri-cious �-V2O5 were detected as soon as the oxidation started,while at high temperatures only AlVO4 and TiO2 were identi-fied [8]. For CrAlVN coatings, AlVO4, (Al, Cr, V)2O3 as well as�-V2O5 oxides were observed for an annealing temperature of700 ◦C [37]. In these studies, the lower onset point of oxidationdisplayed by the V-containing coatings was explained by the reac-tions occurring between protective oxides and vanadium (such asformation of Al–V–O phases). In the present TiSiVN system com-bination of V with titanium or silicon oxide was not detected.Therefore, we suggest that the loss of oxidation resistance ofTiSiVN coatings is due to the rapid oxidation of V and the out-wards diffusion of its ions through the Ti-rich oxide scale. Furtherwork is needed to understand the mechanisms involved duringthe oxidation deterioration of TiSiN coatings with V additionsand their corresponding oxide scale formation. The performanceof solid solution coatings for high temperature sliding applica-tion should be investigated in further work. Rapid formation ofvanadium oxide could indeed reduce friction at temperatureshigher than 500 ◦C; nonetheless, the wear resistance of the coat-ing can be significantly compromised by low oxidation resistance.
In future research, we propose to deposit nanocomposite struc-ture of TiSiVN combining nanograins of c-TiVN embedded intoamorphous Si–N matrix, to control the outwards diffusion ofvanadium.
1 urface
4
tTtAsrtwoptdipssowotVtc
A
tCpCCa
R
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
[
22 F. Fernandes et al. / Applied S
. Conclusion
This investigation concerned the influence of V additions onhe structure, mechanical properties and oxidation resistance ofi–Si–V–N coatings deposited by DC reactive magnetron sput-ering. These coatings were compared to TiN and TiSiN films.ccording to XRD analyses, all coatings showed an fcc NaCl-typetructure assigned to crystalline TiN. A shift of the peaks to theight was observed with Si and V additions, indicative of a substitu-ional solid solution. Hardness and Young modulus of TiSiN coatingsas insignificantly changed with increasing V content. The onset of
xidation of the coatings decreased with V additions down to tem-eratures as low as 500 ◦C, independently of the Si and V content inhe coatings. TiN and TiSiN coating exhibits a typical parabolic oxi-ation weight gain as a function of time, while a different evolution
s displayed by TiSiVN films. At temperatures below the meltingoint of �-V2O5 (∼685 ◦C) two stages were exhibited: at an earlytage, the weight increase over time is linear, whilst, in a secondtage a parabolic evolution could be fitted to the experimental data;n the other hand, at high temperatures only a parabolic evolutionas fitted. �-V2O5 showed to be the main phase present at the
xidized surface of coatings. Reduction of this phase occurred foremperatures above their melting point. The relative amounts of2O5 detected at the oxidized surface of V rich films are promising
o achieve the envisaged good tribological properties; however itan be significantly compromised by their low oxidation resistance.
cknowledgments
This research is sponsored by FEDER funds throughhe program COMPETE–Programa Operacional Factores deompetitividade–and by national funds through FCT – Fundac ãoara a Ciência e a Tecnologia, under the projects: PEst-/EME/UI0285/2013, CENTRO-07-0224-FEDER-002001 (Maisentro SCT 2011 02 001 4637), PTDC/EME-TME/122116/2010nd Plungetec, as well as the grant (SFRH/BD/68740/2010).
eferences
[1] Z. Zhou, W.M. Rainforth, D.B. Lewis, S. Creasy, J.J. Forsyth, F. Clegg, A.P. Ehiasar-ian, P.E. Hovespian, W.D. Münz, Oxidation behaviour of nanoscale TiAlN/VNmultilayer coatings, Surf. Coat. Technol. 177–178 (2004) 198–203.
[2] M. Pfeiler, K. Kutschej, M. Penoy, C. Michotte, C. Mitterer, M. Kathrein, The effectof increasing V content on structure, mechanical and tribological properties ofarc evaporated Ti–Al–V–N coatings, Int. J. Refract. Met. Hard Mater 27 (2009)502–506.
[3] N. Fateh, G.A. Fontalvo, G. Gassner, C. Mitterer, Influence of high-temperatureoxide formation on the tribological behaviour of TiN and VN coatings, Wear262 (2007) 1152–1158.
[4] E. Badisch, G.A. Fontalvo, M. Stoiber, C. Mitterer, Tribological behavior of PACVDTiN coatings in the temperature range up to 500 ◦C, Surf. Coat. Technol. 163–164(2003) 585–590.
[5] M. Stoiber, E. Badisch, C. Lugmair, C. Mitterer, Low-friction TiN coatingsdeposited by PACVD, Surf. Coat. Technol. 163–164 (2003) 451–456.
[6] E. Lugscheider, O. Knotek, K. Bobzin, S. Bärwulf, Tribological properties, phasegeneration and high temperature phase stability of tungsten- and vanadium-oxides deposited by reactive MSIP–PVD process for innovative lubricationapplications, Surf. Coat. Technol. 133–134 (2000) 362–368.
[7] D.B. Lewis, S. Creasey, Z. Zhou, J.J. Forsyth, A.P. Ehiasarian, P.E. Hovsepian, Q.Luo, W.M. Rainforth, W.D. Münz, The effect of (Ti + Al):V ratio on the structureand oxidation behaviour of TiAlN/VN nano-scale multilayer coatings, Surf. Coat.Technol. 177–178 (2004) 252–259.
[8] K. Kutschej, P.H. Mayrhofer, M. Kathrein, P. Polcik, C. Mitterer, Influence ofoxide phase formation on the tribological behaviour of Ti–Al–V–N coatings,Surf. Coat. Technol. 200 (2005) 1731–1737.
[9] K. Kutschej, P.H. Mayrhofer, M. Kathrein, P. Polcik, C. Mitterer, A new low-friction concept for Ti1−xAlxN based coatings in high-temperature applications,Surf. Coat. Technol. 188–189 (2004) 358–363.
10] G. Gassner, P.H. Mayrhofer, K. Kutschej, C. Mitterer, M. Kathrein, A new low
friction concept for high temperatures: lubricious oxide formation on sputteredVN coatings, Tribol. Lett. 17 (2004) 751–756.
12] A. Glaser, S. Surnev, F.P. Netzer, N. Fateh, G.A. Fontalvo, C. Mitterer, Oxida-tion of vanadium nitride and titanium nitride coatings, Surf. Sci. 601 (2007)1153–1159.
13] M. Uchida, N. Nihira, A. Mitsuo, K. Toyoda, K. Kubota, T. Aizawa,Friction and wear properties of CrAlN and CrVN films deposited bycathodic arc ion plating method, Surf. Coat. Technol. 177–178 (2004)627–630.
14] J.H. Ouyang, S. Sasaki, Tribo-oxidation of cathodic arc ion-plated(V,Ti)N coatings sliding against a steel ball under both unlubricatedand boundary-lubricated conditions, Surf. Coat. Technol. 187 (2004)343–357.
15] J.-K. Park, Y.-J. Baik, Increase of hardness and oxidation resistance of VN coat-ing by nanoscale multilayered structurization with AlN, Mater. Lett. 62 (2008)2528–2530.
16] R. Franz, J. Neidhardt, R. Kaindl, B. Sartory, R. Tessadri, M. Lechthaler, P. Polcik,C. Mitterer, Influence of phase transition on the tribological performance of arc-evaporated AlCrVN hard coatings, Surf. Coat. Technol. 203 (2009) 1101–1105.
17] Y. Qiu, S. Zhang, J.-W. Lee, B. Li, Y. Wang, D. Zhao, Self-lubricating CrAlN/VNmultilayer coatings at room temperature, Appl. Surf. Sci. 279 (2013) 189–196.
18] W. Tillmann, S. Momeni, F. Hoffmann, A study of mechanical and tribologicalproperties of self-lubricating TiAlVN coatings at elevated temperatures, Tribol.Int. 66 (2013) 324–329.
19] Q. Luo, Temperature dependent friction and wear of magnetron sputtered coat-ing TiAlN/VN, Wear 271 (2011) 2058–2066.
20] S.H. Kim, J.K. Kim, K.H. Kim, Influence of deposition conditions on themicrostructure and mechanical properties of Ti–Si–N films by DC reactive mag-netron sputtering, Thin Solid Films 420–421 (2002) 360–365.
21] M. Nose, Y. Deguchi, T. Mae, E. Honbo, T. Nagae, K. Nogi, Influence ofsputtering conditions on the structure and properties of Ti–Si–N thin filmsprepared by r.f.-reactive sputtering, Surf. Coat. Technol. 174-175 (2003)261–265.
22] S. Veprek, H.D. Männling, P. Karvankova, J. Prochazka, The issue of thereproducibility of deposition of superhard nanocomposites with hardness of≥50 GPa, Surf. Coat. Technol. 200 (2006) 3876–3885.
23] D. Pilloud, J.F. Pierson, M.C. Marco de Lucas, A. Cavaleiro, Study of the structuralchanges induced by air oxidation in Ti–Si–N hard coatings, Surf. Coat. Technol.202 (2008) 2413–2417.
24] M. Diserens, J. Patscheider, F. Levy, Mechanical properties and oxidation resis-tance of nanocomposite TiN–SiNx physical-vapor-deposited thin films, Surf.Coat. Technol. 120 (1999) 158–165.
25] F. Vaz, L. Rebouta, P. Goudeau, J. Pacaud, H. Garem, J.P. Rivière, A. Cavaleiro, E.Alves, Characterisation of Ti1−xSixNy nanocomposite films, Surf. Coat. Technol.133–134 (2000) 307–313.
26] M. Diserens, J. Patscheider, F. Lévy, Improving the properties of titanium nitrideby incorporation of silicon, Surf. Coat. Technol. 108–109 (1998) 241–246.
27] Y.H. Cheng, T. Browne, B. Heckerman, E.I. Meletis, Mechanical and tribologicalproperties of nanocomposite TiSiN coatings, Surf. Coat. Technol. 204 (2010)2123–2129.
28] J. Patscheider, T. Zehnder, M. Diserens, Structure–performance relations innanocomposite coatings, Surf. Coat. Technol. 146–147 (2001) 201–208.
29] D. Ma, S. Ma, K. Xu, The tribological and structural characterization of nano-structured Ti–Si–N films coated by pulsed-d.c. plasma enhanced CVD, Vacuum79 (2005) 7–13.
30] G. Stoney, The tension of metallic films deposited by electrolysis, Proc. R. Soc.London, Ser. A 82 (1909) 172–175.
31] C.P. Constable, D.B. Lewis, J. Yarwood, W.D. Münz, Raman microscopic studiesof residual and applied stress in PVD hard ceramic coatings and correlationwith X-ray diffraction (XRD) measurements, Surf. Coat. Technol. 184 (2004)291–297.
32] F. Vaz, L. Rebouta, B. Almeida, P. Goudeau, J. Pacaud, J.P. Rivière, J. Bessa e Sousa,Structural analysis of Ti1−xSixNy nanocomposite films prepared by reactivemagnetron sputtering, Surf. Coat. Technol. 120–121 (1999) 166–172.
33] M. Diserens, J. Patscheider, F. Lévy, Mechanical properties and oxidation resis-tance of nanocomposite TiN–SiNx physical-vapor-deposited thin films, Surf.Coat. Technol. 120–121 (1999) 158–165.
34] T. Kacsich, S. Gasser, Y. Tsuji, A. Dommann, M.A. Nicolet, A. Nicolet, Wet oxida-tion of Ti34Si23B43, Br. J. Appl. Phys. 85 (1999) 1871–1875.
35] T. Kacsich, M.A. Nicolet, Moving species in Ti34Si23N43 oxidation, Thin SolidFilms 349 (1999) 1–3.
36] Z. Zhou, W.M. Rainforth, C. Rodenburg, N.C. Hyatt, D.B. Lewis, P.E. Hovsepian,Oxidation Behavior and mechanisms of TiAlN/VN coatings, Metall. Mater. Trans.A 38 (2007) 2464–2478.
37] R. Franz, J. Neidhardt, C. Mitterer, B. Schaffer, H. Hutter, R. Kaindl, B. Sartory,R. Tessadri, M. Lechthaler, P. Polcik, Oxidation and diffusion processes duringannealing of AlCrVN hard coatings, J. Vac. Sci. Technol., A 26 (2008) 302–308.
38] M. Wittmer, J. Noser, H. Melchior, Oxidation kinetics of TiN thin films, J. Appl.Phys. 52 (1981) 6659–6664.
39] C.W. Zou, X.D. Yan, J. Han, R.Q. Chen, W. Gao, Microstructures and optical prop-erties of �-V 2 O 5 nanorods prepared by magnetron sputtering, J. Phys. D: Appl.Phys. 42 (2009) 145402.
40] A. Bouzidi, N. Benramdane, S. Bresson, C. Mathieu, R. Desfeux, M.E. Marssi, X-ray
and Raman study of spray pyrolysed vanadium oxide thin films, Vib. Spectrosc.57 (2011) 182–186.
41] M. Fernández-García, X. Wang, C. Belver, J.C. Hanson, J.A. Rodriguez, Anatase-TiO2 nanomaterials: morphological/size dependence of the crystallization andphase behavior phenomena, J. Phys. Chem. C 111 (2006) 674–682.